Дисертації: "HYPER SPECTRAL IMAGE CLASSIFICATION"

1

Kliman, Douglas Hartley. "Rule-based classification of hyper-temporal, multi-spectral satellite imagery for land-cover mapping and monitoring." Diss., The University of Arizona, 1996. http://hdl.handle.net/10150/187473.

Повний текст джерела

Анотація:

A rule-based classification model was developed to derive land-cover information from a large set of hyper-temporal, multi-spectral satellite imagery encompassing the state of Arizona. The model uses Advanced Very High Resolution Radiometer (AVHRR) imagery and the 30-minute digital elevation model (DEM) from the EROS Data Center (EDC) Conterminous U.S. AVHRR Biweekly Composites. Sixty one images from 1990, 1991 and 1992 were analyzed using the Brown & Lowe (1973) Natural Vegetative Communities of Arizona map to identify temporal patterns of Normalized Difference Vegetation Index (NDVI) and thermal measurements for 13 land-cover classes. Fifteen characteristic layers were created to represent the spectral, thermal and temporal properties of the data set. These layers were inputs for the rule-based classification model. The model was run on three years of data, creating three single year land-cover maps. The modeling effort showed that NDVI, thermal and DEM characteristics are useful for discerning land-cover classes. The single year land-cover maps showed that the rule-based model could not detect land-cover change between years. The single year maps were combined to create a summary land-cover map. This map differs from the Brown and Lowe map in the shape, proportional size and spatial distribution of land-cover polygons. The rule-based model can discern more land-cover classes than spectral cluster classification. Ground observations and an aerial video was used to assess map accuracy. The same proportion of agreement was observed between the ground observations, the Brown and Lowe map, and the summary land-cover map. Agreement was higher between video and the summary map than between video and the Brown and Lowe map. With further refinements to the input data set, classification model rules and field accuracy assessment, higher levels of agreement can be expected. Overall results show that rule-based classification of hyper-temporal, multi-spectral satellite imagery is a desirable method for mapping global land-cover.

Стилі APA, Harvard, Vancouver, ISO та ін.

2

Falco, Nicola. "Advanced Spectral and Spatial Techniques for Hyperspectral Image Analysis and Classification." Doctoral thesis, Università degli studi di Trento, 2015. https://hdl.handle.net/11572/369072.

Повний текст джерела

Анотація:

Recent advances in sensor technology have led to an increased availability of hyperspectral remote sensing images with high spectral and spatial resolutions. These images are composed by hundreds of contiguous spectral channels, covering a wide spectral range of frequencies, in which each pixel contains a highly detailed representation of the reflectance of the materials present on the ground, and a better characterization in terms of geometrical detail. The burst of informative content conveyed in the hyperspectral images permits an improved characterization of different land coverages. In spite of that, it increases significantly the complexity of the analysis, introducing a series of challenges that need to be addressed, such as the computational complexity and resources required. This dissertation aims at defining novel strategies for the analysis and classification of hyperspectral remote sensing images, placing the focal point on the investigation and optimisation techniques for the extraction and integration of spectral and spatial information. In the first part of the thesis, a thorough study on the analysis of the spectral information contained in the hyperspectral images is presented. Though, independent component analysis (ICA) has been widely used to address several tasks in the remote sensing field, such as feature reduction, spectral unmixing and classification, its employment in extracting class-discriminant information remains a research topic open to further investigation. To this extend, a profound study on the performances of different ICA algorithms is performed, highlighting their strengths and weaknesses in the hyperspectral image classification task. Based on this study, a novel approach for feature reduction is proposed, where the use of ICA is optimised for the extraction of class-specific information. In the second part of the thesis, the spatial information is exploited by employing operators from the mathematical morphology framework. Morphological operators, such as attribute profiles and their multi-channel and multi-attribute extensions, are proved to be effective in the modelling of the spatial information, dealing, however, with issues such as the high feature dimensionality, the high intrinsic information redundancy and the a-priori need for parameter tuning in filtering, which are still open. Addressing the first two issues, the reduced attribute profiles are introduced, in this thesis, as an optimised version of the morphological attribute profiles, with the property to compress all the meaningful geometrical information into a few features. Regarding the filter parameter tuning issue, an innovative strategy for automatic threshold selection is proposed. Inspired by the concept of granulometry, the proposed approach defines a novel granulometric characteristic function, which provides information on the image decomposition according to a given measure. The approach exploits the tree representation of an image, allowing us to avoid additional filtering steps prior to the threshold selection, making the process computationally effective. The outcome of this dissertation advances the state-of-the-art by proposing novel methodologies for accurate hyperspectral image classification, where the results obtained by extensive experimentation on various real hyperspectral data sets confirmed their effectiveness. Concluding the thesis, insightful and concrete remarks to the aforementioned issues are discussed.

Стилі APA, Harvard, Vancouver, ISO та ін.

3

Falco, Nicola. "Advanced Spectral and Spatial Techniques for Hyperspectral Image Analysis and Classification." Doctoral thesis, University of Trento, 2015. http://eprints-phd.biblio.unitn.it/1421/1/PhD_Nicola_Trento.pdf.

Повний текст джерела

Анотація:

Recent advances in sensor technology have led to an increased availability of hyperspectral remote sensing images with high spectral and spatial resolutions. These images are composed by hundreds of contiguous spectral channels, covering a wide spectral range of frequencies, in which each pixel contains a highly detailed representation of the reflectance of the materials present on the ground, and a better characterization in terms of geometrical detail. The burst of informative content conveyed in the hyperspectral images permits an improved characterization of different land coverages. In spite of that, it increases significantly the complexity of the analysis, introducing a series of challenges that need to be addressed, such as the computational complexity and resources required. This dissertation aims at defining novel strategies for the analysis and classification of hyperspectral remote sensing images, placing the focal point on the investigation and optimisation techniques for the extraction and integration of spectral and spatial information. In the first part of the thesis, a thorough study on the analysis of the spectral information contained in the hyperspectral images is presented. Though, independent component analysis (ICA) has been widely used to address several tasks in the remote sensing field, such as feature reduction, spectral unmixing and classification, its employment in extracting class-discriminant information remains a research topic open to further investigation. To this extend, a profound study on the performances of different ICA algorithms is performed, highlighting their strengths and weaknesses in the hyperspectral image classification task. Based on this study, a novel approach for feature reduction is proposed, where the use of ICA is optimised for the extraction of class-specific information. In the second part of the thesis, the spatial information is exploited by employing operators from the mathematical morphology framework. Morphological operators, such as attribute profiles and their multi-channel and multi-attribute extensions, are proved to be effective in the modelling of the spatial information, dealing, however, with issues such as the high feature dimensionality, the high intrinsic information redundancy and the a-priori need for parameter tuning in filtering, which are still open. Addressing the first two issues, the reduced attribute profiles are introduced, in this thesis, as an optimised version of the morphological attribute profiles, with the property to compress all the meaningful geometrical information into a few features. Regarding the filter parameter tuning issue, an innovative strategy for automatic threshold selection is proposed. Inspired by the concept of granulometry, the proposed approach defines a novel granulometric characteristic function, which provides information on the image decomposition according to a given measure. The approach exploits the tree representation of an image, allowing us to avoid additional filtering steps prior to the threshold selection, making the process computationally effective. The outcome of this dissertation advances the state-of-the-art by proposing novel methodologies for accurate hyperspectral image classification, where the results obtained by extensive experimentation on various real hyperspectral data sets confirmed their effectiveness. Concluding the thesis, insightful and concrete remarks to the aforementioned issues are discussed.

Стилі APA, Harvard, Vancouver, ISO та ін.

4

Jia, Xiuping Electrical Engineering Australian Defence Force Academy UNSW. "Classification techniques for hyperspectral remote sensing image data." Awarded by:University of New South Wales - Australian Defence Force Academy. School of Electrical Engineering, 1996. http://handle.unsw.edu.au/1959.4/38713.

Повний текст джерела

Анотація:

Hyperspectral remote sensing image data, such as that recorded by AVIRIS with 224 spectral bands, provides rich information on ground cover types. However, it presents new problems in machine assisted interpretation, mainly in long processing times and the difficulties of class training due to the low ratio of number of training samples to the number of bands. This thesis investigates feasible and efficient feature reduction and image classification techniques which are appropriate for hyperspectral image data. The study is reported in three parts. The first concerns a deterministic approach for hyperspectral data interpretation. Multigroup and multiple threshold spectral coding procedures, and associated techniques for spectral matching and classification, are proposed and tested. By coding on subgroups of bands using one or three thresholds, spectral searching and matching becomes simple, fast and free of the need for radiometric correction. Modifications of existing statistical techniques are proposed in the second part of the investigation A block-based maximum likelihood classification technique is developed. Several subgroups are formed from the complete set of spectral bands in the data, based on the properties of global correlation among the bands. Subgroups which are poorly correlated with each other are treated independently using conventional maximum likelihood classification. Experimental results demonstrate that, when using appropriate subgroup sizes, the new method provides a compromise among classification accuracy, processing time and available training pixels. Furthermore, a segmented, and possibly multi-layer, principal components transformation is proposed as a possible feature reduction technique prior to classification, and for effective colour display. The transformation is performed efficiently on each of the highly correlated subgroups of bands independently. Selected features from each transformed subgroup can be then transformed again to achieve a satisfactory data reduction ratio and to generate the three most significant components for colour display. Classification accuracy is improved and high quality colour image display is achieved in experiments using two AVIRIS data sets.

Стилі APA, Harvard, Vancouver, ISO та ін.

5

Prasert, Sunyaruk. "Multi angle imaging with spectral remote sensing for scene classification." Thesis, Monterey, Calif. : Springfield, Va. : Naval Postgraduate School ; Available from National Technical Information Service, 2005. http://library.nps.navy.mil/uhtbin/hyperion/05Mar%5FPrasert.pdf.

Повний текст джерела

Анотація:

Thesis (M.S. in Information Technology Management)--Naval Postgraduate School, March 2005.
Thesis Advisor(s): Richard C. Olsen. Includes bibliographical references (p. 95-97). Also available online.

Стилі APA, Harvard, Vancouver, ISO та ін.

6

Alam, Fahim Irfan. "Deep Feature Learning for Spectral-Spatial Classification of Hyperspectral Remote Sensing Images." Thesis, Griffith University, 2019. http://hdl.handle.net/10072/386535.

Повний текст джерела

Анотація:

The recent advances in aerial- and satellite-based hyperspectral imaging sensor technologies have led to an increased availability of Earth's images with high spatial and spectral resolution, which opened the door to a large range of important applications. Hyperspectral imaging records detailed spectrum of the received light in each spatial position in the image, in which each pixel contains a highly detailed representation of the reflectance of the materials present on the ground, and a better characterization in terms of geometrical details. Since different substances exhibit different spectral signatures, the abundance of informative content conveyed in the hyperspectral images permits an improved characterization of different land coverage. Therefore, hyperspectral imaging emerged as a well-suited technology for accurate image classi fication in remote sensing. In spite of that, a signi ficantly increased complexity of the analysis introduces a series of challenges that need to be addressed on a serious note. In order to fully exploit the potential offered by these sensors, there is a need to develop accurate and effective models for spectral-spatial analysis of the recorded data. This thesis aims at presenting novel strategies for the analysis and classifi cation of hyperspectral remote sensing images, placing the focal point on the investigation on deep networks for the extraction and integration of spectral and spatial information. Deep learning has demonstrated cutting-edge performances in computer vision, particularly in object recognition and classi cation. It has also been successfully adopted in hyperspectral remote sensing domain as well. However, it is a very challenging task to fully utilize the massive potential of deep models in hyperspectral remote sensing applications since the number of training samples is limited which limits the representation capability of a deep model. Furthermore, the existing architectures of deep models need to be further investigated and modifi ed accordingly to better complement the joint use of spectral and spatial contents of hyperspectral images. In this thesis, we propose three different deep learning-based models to effectively represent spectral-spatial characteristics of hyperspectral data in the interest of classifi cation of remote sensing images. Our first proposed model focuses on integrating CRF and CNN into an end-to-end learning framework for classifying images. Our main contribution in this model is the introduction of a deep CRF in which the CRF parameters are computed using CNN and further optimized by adopting piecewise training. Furthermore, we address the problem of over fitting by employing data augmentation techniques and increased the size of the training samples for training deep networks. Our proposed 3DCNN-CRF model can be trained to fully exploit the usefulness of CRF in the context of classi fication by integrating it completely inside of a deep model. Considering that the separation of constituent materials and their abundances provide detailed analysis of the data, our second algorithm investigates the potential of using unmixing results in deep models to classify images. We extend an existing region based structure preserving non-negative matrix factorization method to estimate groups of spectral bands with the goal to capture subtle spectral-spatial distribution from the image. We subsequently use these important unmixing results as input to generate superpixels, which are further represented by kernel density estimated probability distribution function. Finally, these abundance information-guided superpixels are directly supplied into a deep model in which the inference is implicitly formulated as a recurrent neural network to perform the eventual classifi cation. Finally, we perform a detailed investigation on the possibilities of adopting generative adversarial models into hyperspectral image classifi cation. We present a GAN-based spectral-spatial method that primarily focuses on signifi cantly improving the multiclass classi cation ability of the discriminator of GAN models. In this context, we propose to adopt the triplet constraint property and extend it to build a useful feature embedding for remote sensing images for use in classi cation. Furthermore, our proposed Triplet- 3D-GAN model also includes feedback from discriminator's intermediate features to improve the quality of the generator's sample generation process.
Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
School of Info & Comm Tech
Science, Environment, Engineering and Technology
Full Text

Стилі APA, Harvard, Vancouver, ISO та ін.

7

Hoarau, Romain. "Rendu interactif d'image hyper spectrale par illumination globale pour la prédiction de la signature infrarouge d'aéronefs." Electronic Thesis or Diss., Aix-Marseille, 2019. http://theses.univ-amu.fr.lama.univ-amu.fr/191219_HOARAU_358wfqq893efe918esmfu405fjhqvj_TH.pdf.

Повний текст джерела

Анотація:

Le dimensionnement de capteur est un enjeu majeur pour le domaine de la détection d'aéronefs. Dans cette optique, il est nécessaire de simuler ces capteurs via des modèles et un nombre conséquent d'images spectrales d'aéronefs. L'obtention de ces images via des campagnes aériennes de mesure est toutefois onéreuse et difficile. Une simulation de ces données s'impose donc. Afin de répondre à ces besoins, des algorithmes d'illumination globale à haute dimension spectrale sont utilisés. Dans ces conditions, ces algorithmes posent des problèmes de consommation mémoire et de temps de calcul. Le projet de recherche de cette thèse s'inscrit dans le cadre de ces problématiques.Dans un premier temps, nous nous sommes focalisés sur l'algorithme du Path Tracing et la parallélisation GPUpour le rendu d'images spectrales. Nous avons d'abord analysé les problèmes de ce type de rendu sur GPU.Nous avons ensuite proposé une nouvelle méthode et un schéma de parallélisation spectral qui permettent de réduire significativement la consommation mémoire et les temps de calcul.Dans un second temps, nous avons cherché à réduire la charge de calcul spectrale de la simulation. À cet égard, nous avons proposé de généraliser le rendu spectral stochastique d'image dans l'espace CIE XYZ en rendu d'image spectrale stochastique. Cette méthode permet de rendre directement et de manière plus précise et rapide les canaux d'un capteur en diminuant la dimension spectrale de la simulation. Pour conclure, les travaux de cette thèse permettent de simuler de manière précise des images multi, hyper et ultra spectrales. Le temps interactif peut être atteint dans notre cas en multi et hyper spectrale
Sensor dimensioning is a major issue for the aircraft detection field. In this vein, it is appropriate to simulate these sensorsvia models and a consequent set of spectral images. The acquisition of these images via an airborne measure campaign is unfortunately costly and difficult. A robust and fast simulation of these data is hence very appealing.In order to answer these needs, global illumination methods in high spectral dimension are used. In these circumstances,these methods raise serious issues in term of memory consumption and of computing time. Our research project focuses on these problematics.In the first instance, we have focused on the Path Tracing method and its GPU parallelization for the spectral image rendering. We have investigated at first the issues of this kind of rendering on the GPU. Then we have proposed a new method and an efficient spectral parallelization pattern which allows us to reduce significantly the memory consumption and thecomputing time.In the second phase, we have investigated how to reduce the spectral computational load of the simulation. Inthat sense, we have proposed to generalize the stochastic spectral rendering of color (XYZ) image to the stochastic spectral image rendering. This new method renders directly the channels of a sensor which allows us to reduce the memory andthe computing requirements by reducing the spectral computational load of the simulation.To sum up, the works of this thesis allows us to simulate accurately multi, hyper and ultra spectral images. The interactive time can be achieved in our case in multi and hyper spectral resolution

Стилі APA, Harvard, Vancouver, ISO та ін.

8

Behmo, Régis. "Visual feature graphs and image recognition." Phd thesis, Ecole Centrale Paris, 2010. http://tel.archives-ouvertes.fr/tel-00545419.

Повний текст джерела

Анотація:

La problèmatique dont nous nous occupons dans cette thèse est la classification automatique d'images bidimensionnelles, ainsi que la détection d'objets génériques dans des images. Les avancées de ce champ de recherche contribuent à l'élaboration de systèmes intelligents, tels que des robots autonomes et la création d'un web sémantique. Dans ce contexte, la conception de représentations d'images et de classificateurs appropriés constituent des problèmes ambitieux. Notre travail de recherche fournit des solutions à ces deux problèmes, que sont la représentation et la classification d'images. Afin de générer notre représentation d'image, nous extrayons des attributs visuels de l'image et construisons une structure de graphe basée sur les propriétés liées au relations de proximités entre les points d'intérêt associés. Nous montrons que certaines propriétés spectrales de ces graphes constituent de bons invariants aux classes de transformations géométriques rigides. Notre représentation d'image est basée sur ces propriétés. Les résultats expérimentaux démontrent que cette représentation constitue une amélioration par rapport à d'autres représentations similaires, mais qui n'intègrent pas les informations liées à l'organisation spatiale des points d'intérêt. Cependant, un inconvénient de cette méthode est qu'elle fait appel à une quantification (avec pertes) de l'espace des attributs visuels afin d'être combinée avec un classificateur Support Vecteur Machine (SVM) efficace. Nous résolvons ce problème en créant un nouveau classificateur, basé sur la distance au plus proche voisin, et qui permet la classification d'objets assimilés à des ensembles de points. La linéarité de ce classificateur nous permet également de faire de la détection d'objet, en plus de la classification d'images. Une autre propriété intéressante de ce classificateur est sa capacité à combiner différents types d'attributs visuels de manière optimale. Nous utilisons cette propriété pour formuler le problème de classification de graphes de manière différente. Les expériences, menées sur une grande variété de jeux de données, montrent les bénéfices quantitatifs de notre approche.

Стилі APA, Harvard, Vancouver, ISO та ін.

9

Tso, Brandt C. K. "An investigation of alternative strategies for incorporating spectral, textural, and contextual information in remote sensing image classification." Thesis, University of Nottingham, 1997. http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.387663.

Повний текст джерела

Стилі APA, Harvard, Vancouver, ISO та ін.

10

Rajadell, Rojas Olga. "Data selection and spectral-spatial characterisation for hyperspectral image segmentation. Applications to remote sensing." Doctoral thesis, Universitat Jaume I, 2013. http://hdl.handle.net/10803/669093.

Повний текст джерела

Анотація:

El análisis de imágenes ha impulsado muchos descubrimientos en la ciencia actual. Esta tesis se centra en el análisis de imágenes remotas para inspección aérea, exactamente en el problema de segmentación y clasificación de acuerdo al uso del suelo. Desde el nacimiento de los sensores hiperespectrales su uso ha sido vital para esta tarea ya que facilitan y mejoran sustancialmente el resultado. Sin embargo el uso de imágenes hiperespectrales entraña, entre otros, problemas de dimensionalidad y de interacción con los expertos. Proponemos mejoras que ayuden a paliar estos inconvenientes y hagan el problema mas eficiente.
Lately image analysis have aided many discoveries in research. This thesis focusses on the analysis of remote sensed images for aerial inspection. It tackles the problem of segmentation and classification according to land usage. In this field, the use of hyperspectral images has been the trend followed since the emergence of hyperspectral sensors. This type of images improves the performance of the task but raises some issues. Two of those issues are the dimensionality and the interaction with experts. We propose enhancements overcome them. Efficiency and economic reasons encouraged to start this work. The enhancements introduced in this work allow to tackle segmentation and classification of this type of images using less data, thus increasing the efficiency and enabling the design task specific sensors which are cheaper. Also, our enhacements allow to perform the same task with less expert collaboration which also decreases the costs and accelerates the process.

Стилі APA, Harvard, Vancouver, ISO та ін.

11

Teresi, Michael Bryan. "Multispectral Image Labeling for Unmanned Ground Vehicle Environments." Thesis, Virginia Tech, 2015. http://hdl.handle.net/10919/53998.

Повний текст джерела

Анотація:

Described is the development of a multispectral image labeling system with emphasis on Unmanned Ground Vehicles(UGVs). UGVs operating in unstructured environments face significant problems detecting viable paths when LIDAR is the sole source for perception. Promising advances in computer vision and machine learning has shown that multispectral imagery can be effective at detecting materials in unstructured environments [1][2][3][4][5][6]. This thesis seeks to extend previous work[6][7] by performing pixel level classification with multispectral features and texture. First the images are spatially registered to create a multispectral image cube. Visual, near infrared, shortwave infrared, and visible/near infrared polarimetric data are considered. The aligned images are then used to extract features which are fed to machine learning algorithms. The class list includes common materials present in rural and urban scenes such as vehicles, standing water, various forms of vegetation, and concrete. Experiments are conducted to explore the data requirement for a desired performance and the selection of a hyper-parameter for the textural features. A complete system is demonstrated, progressing from the data collection and labeling to the analysis of the classifier performance.
Master of Science

Стилі APA, Harvard, Vancouver, ISO та ін.

12

Key, Thomas Lee. "An evaluation of the relative value of spectral and phenological information for tree crown classification of digital images in the eastern deciduous forest /." Morgantown, W. Va. : [West Virginia University Libraries], 1998. http://etd.wvu.edu/templates/showETD.cfm?recnum=107.

Повний текст джерела

Анотація:

Thesis (M.A.)--West Virginia University, 1998.
Title from document title page. Document formatted into pages; contains viii, 51 p. : col. ill., col. map. Vita. Includes abstract. Includes bibliographical references (p. 32-34).

Стилі APA, Harvard, Vancouver, ISO та ін.

13

Wondim, Yonas kassaw. "Hyperspectral Image Analysis Algorithm for Characterizing Human Tissue." Thesis, Linköpings universitet, Biomedicinsk instrumentteknik, 2011. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-75156.

Повний текст джерела

Анотація:

AbstractIn the field of Biomedical Optics measurement of tissue optical properties, like absorption, scattering, and reduced scattering coefficient, has gained importance for therapeutic and diagnostic applications. Accuracy in determining the optical properties is of vital importance to quantitatively determine chromophores in tissue.There are different techniques used to quantify tissue chromophores. Reflectance spectroscopy is one of the most common methods to rapidly and accurately characterize the blood amount and oxygen saturation in the microcirculation. With a hyper spectral imaging (HSI) device it is possible to capture images with spectral information that depends both on tissue absorption and scattering. To analyze this data software that accounts for both absorption and scattering event needs to be developed.In this thesis work an HSI algorithm, capable of assessing tissue oxygenation while accounting for both tissue absorption and scattering, is developed. The complete imaging system comprises: a light source, a liquid crystal tunable filter (LCTF), a camera lens, a CCD camera, control units and power supply for light source and filter, and a computer.This work also presents a Graphic processing Unit (GPU) implementation of the developed HSI algorithm, which is found computationally demanding. It is found that the GPU implementation outperforms the Matlab “lsqnonneg” function by the order of 5-7X.At the end, the HSI system and the developed algorithm is evaluated in two experiments. In the first experiment the concentration of chromophores is assessed while occluding the finger tip. In the second experiment the skin is provoked by UV light while checking for Erythema development by analyzing the oxyhemoglobin image at different point of time. In this experiment the melanin concentration change is also checked at different point of time from exposure.It is found that the result matches the theory in the time dependent change of oxyhemoglobin and deoxyhemoglobin. However, the result of melanin does not correspond to the theoretically expected result.

Стилі APA, Harvard, Vancouver, ISO та ін.

14

Hijazi, Hala. "Proposition d'une méthode spectrale combinée LDA et LLE pour la réduction non-linéaire de dimension : Application à la segmentation d'images couleurs." Thesis, Littoral, 2013. http://www.theses.fr/2013DUNK0516.

Повний текст джерела

Анотація:

Les méthodes d'analyse de données et d'apprentissage ont connu un développement très important ces dernières années. En effet, après les réseaux de neurones, les machines à noyaux (années 1990), les années 2000 ont vu l'apparition de méthodes spectrales qui ont fourni un cadre mathématique unifié pour développer des méthodes de classification originales. Parmi celles-ci ont peut citer la méthode LLE pour la réduction de dimension non linéaire et la méthode LDA pour la discrimination de classes. Une nouvelle méthode de classification est proposée dans cette thèse, méthode issue d'une combinaison des méthodes LLE et LDA. Cette méthode a donné des résultats intéressants sur des ensembles de données synthétiques. Elle permet une réduction de dimension non-linéaire suivie d'une discrimination efficace. Ensuite nous avons montré que cette méthode pouvait être étendue à l'apprentissage semi-supervisé. Les propriétés de réduction de dimension et de discrimination de cette nouvelle méthode, ainsi que la propriété de parcimonie inhérente à la méthode LLE nous ont permis de l'appliquer à la segmentation d'images couleur avec succès. La propriété d'apprentissage semi-supervisé nous a enfin permis de segmenter des images bruitées avec de bonnes performances. Ces résultats doivent être confortés mais nous pouvons d'ores et déjà dégager des perspectives de poursuite de travaux intéressantes
Data analysis and learning methods have known a huge development during these last years. Indeed, after neural networks, kernel methods in the 90', spectral methods appeared in the years 2000. Spectral methods provide an unified mathematical framework to expand new original classification methods. Among these new techniques, two methods can be highlighted : LLE for non-linear dimension reduction and LDA as discriminating classification method. In this thesis document a new classification technique is proposed combining LLE and LDA methods. This new method makes it possible to provide efficient non-linear dimension reduction and discrimination. Then an extension of the method to semi-supervised learning is proposed. Good properties of dimension reduction and discrimination associated with the sparsity property of the LLE technique make it possible to apply our method to color images segmentation with success. Semi-supervised version of our method leads to efficient segmentation of noisy color images. These results have to be extended and compared with other state-of-the-art methods. Nevertheless interesting perspectives of this work are proposed in conclusion for future developments

Стилі APA, Harvard, Vancouver, ISO та ін.

15

Hamed, Nabil. "Conception et realisation d'un systeme de classification en teledetection par combinaison d'analyses radiometriques et spatiales." Université Louis Pasteur (Strasbourg) (1971-2008), 1987. http://www.theses.fr/1987STR13153.

Повний текст джерела

Анотація:

Lorsque la resolution spatiale devient trop fine, l'utilisation de l'analyse radiometrique pour l'extraction d'informations des images en teledetection est insuffisante. Aussi, pour pallier a ce probleme, on presente une nouvelle classification combinant une analyse spatiale a l'analyse radiometrique

Стилі APA, Harvard, Vancouver, ISO та ін.

16

Green, Christopher Lee. "IP Algorithm Applied to Proteomics Data." Diss., CLICK HERE for online access, 2004. http://contentdm.lib.byu.edu/ETD/image/etd618.pdf.

Повний текст джерела

Стилі APA, Harvard, Vancouver, ISO та ін.

17

Linn, Rodrigo de Marsillac. "Avaliação dos modelos de mistura espectral MESMA e SMA aplicados aos dados hiperespectrais Hyperion/EO-1 adquiridos na Planície Costeira do Rio Grande do Sul." reponame:Biblioteca Digital de Teses e Dissertações da UFRGS, 2008. http://hdl.handle.net/10183/23707.

Повний текст джерела

Анотація:

O objetivo do presente trabalho foi avaliar o uso potencial dos dados hiperespectrais do sensor orbital Hyperion/Earth Observing One (EO-1) e dos modelos de mistura espectral MESMA (Multiple Endmember Spectral Mixture Analysis) e SMA (Spectral Mixture Analysis) para discriminação de classes de cobertura da Planície Costeira do Rio Grande do Sul. O modelo MESMA difere do SMA por permitir que o número e o tipo de Membros de Referência (MRs), assim como sua abundância, variem pixel a pixel. A abordagem metodológica utilizada envolveu as seguintes etapas: (a) préprocessamento dos dados Hyperion e conversão dos valores de radiância para imagens atmosfericamente corrigidas de reflectância de superfície; (b) uso seqüencial das técnicas Minimum Noise Fraction (MNF), Pixel Purity Index (PPI) e Visualizador n- Dimensional, no intervalo de 454 a 2334 nm, para seleção inicial de um grupo de pixels candidatos a MRs (primeira biblioteca espectral) e de um outro grupo para fins de validação dos modelos; (c) uso do aplicativo VIPER (Visualization and Image Processing for Environmental Research) Tools para refinamento da primeira biblioteca espectral e seleção final dos MRs, utilizando as métricas EAR (Endmember Average RMSE), MASA (Minimum Average Spectral Angle) e CoB (Count Based Endmember Selection); (d) geração dos modelos MESMA e SMA com o VIPER Tools; e (e) comparação dos resultados dos modelos com base nas imagens-fração e nos valores de erro médio quadrático (RMSE). Os resultados obtidos mostraram que: (1) o uso seqüencial das técnicas MNF, PPI e Visualizador n-Dimensional pode constituir uma etapa inicial para identificar pixels candidatos a MRs, cuja seleção final pode ser feita com as métricas EAR, MASA e CoB. Usadas de forma combinada, essas métricas minimizam possíveis efeitos da baixa relação sinal-ruído do Hyperion; (2) os MRs selecionados representaram os principais componentes de cena como “água” (com clorofila, límpida e com sedimentos em suspensão), “vegetação verde” (pinus, eucalipto e gramíneas) e “solo” (dunas e campo seco); (3) Por utilizar número e tipo variáveis de MRs, o modelo MESMA produziu melhores resultados que o SMA. Quando aplicado sobre a imagem, sobre a amostra de validação e quando comparado com o SMA, o modelo MESMA de 4 componentes (Solo = dunas e campo Seco; vegetação verde = pinus, eucalipto e gramíneas; água = com Sedimentos em suspensão, sem Sedimentos e com clorofila; sombra) descreveu adequadamente a diversidade dos componentes de cena, incluindo materiais dentro de uma mesma classe (p.ex. pinus e eucalipto). O MESMA produziu menores valores de RMSE e uma maior quantidade de pixels modelados na cena (85% contra 55%) do que o SMA; (4) o VIPER mostrou-se uma ferramenta bastante eficaz para seleção dos MRs e geração dos modelos. Os resultados, como um todo, demonstraram o potencial da aplicação dos modelos MESMA com dados hiperespectrais do sensor Hyperion/EO-1, mesmo considerando a baixa relação sinal/ruído do instrumento, especialmente no infravermelho de ondas curtas (SWIR).
The objective of this work was to evaluate the potential use of the Hyperion/Earth Observing One (EO-1) hyperspectral data and of the MESMA (Multiple Endmember Spectral Mixture Analysis) and SMA (Spectral Mixture Analysis) mixture models to discriminate land covers in the Rio Grande do Sul state, South Brazil. MESMA differs from SMA because it may use a variable number and type of endmembers in each pixel. The methodology involved: (a) pre-processing of Hyperion data and conversion of radiance values into atmospherically corrected surface reflectance images; (b) sequential use of the Minimum Noise Fraction (MNF), Pixel Purity Index (PPI) and n- Dimensional Visualizer techniques, in the 454-2334 nm range, for initial selection of a general group of candidate endmembers (first spectral library) and of another group of pixels used for model validation; (c) use of VIPER (Visualization and Image Processing for Environmental Research) Tools algorithm for final selection of endmembers from the first spectral library and from the use of the metrics EAR (Endmember Average RMSE), MASA (Minimum Average Spectral Angle) and CoB (Count Based Endmember Selection); (d) use of VIPER tools to obtain MESMA and SMA models; and (e) comparison of modeling results based on the inspection of fraction images and root mean square error (RMSE) values. Results showed that: (1) the sequential use of the MNF, PPI and n-D Visualizer techniques may comprise an initial step to identify candidate endmembers. Final selection was performed using a combination of EAR, MASA and CoB to minimize possible effects of low signalnoise ratio (SNR) of Hyperion; (2) the selected endmembers represented major scene components such as water (with chlorophyll, clear or bearing in suspended sediments), green vegetation (pinus, eucalyptus and grasslands) and soil (dunes and dry grasslands); (3) By using a variable number and type of endmembers, MESMA produced better results than SMA. When applied over the image, the validation dataset and compared with SMA, the four-endmember MESMA model (soil = dunes and dry grasslands; green vegetation = pinus, eucalyptus and grasslands; water = with chlorophyll, clear and with suspended sediments; shadow) described adequately the diversity of the scene components, including materials within the same class (e.g., pinus and eucalyptus). MESMA produced lower RMSE values and greater number of modeled pixels (85% versus 55%) than SMA; (5) the VIPER tools seems to be an interesting approach for endmember selection and spectral mixture model generation. Results, as a whole, demonstrated the potential use of the MESMA with Hyperion/EO-1 hyperspectral data, even considering the low SNR of the instrument, especially in the shortwave infrared (SWIR).

Стилі APA, Harvard, Vancouver, ISO та ін.

18

Karvir, Hrishikesh. "Design and Validation of a Sensor Integration and Feature Fusion Test-Bed for Image-Based Pattern Recognition Applications." Wright State University / OhioLINK, 2010. http://rave.ohiolink.edu/etdc/view?acc_num=wright1291753291.

Повний текст джерела

Стилі APA, Harvard, Vancouver, ISO та ін.

19

Teixeira, Karla dos Santos. "Uma proposta metodológica de integração de técnicas de análise espectral e de inteligência computacional, baseadas em conhecimento, para o reconhecimento de padrões em imagens multiespectrais." Universidade do Estado do Rio de Janeiro, 2012. http://www.bdtd.uerj.br/tde_busca/arquivo.php?codArquivo=5779.

Повний текст джерела

Анотація:

Somente no ano de 2011 foram adquiridos mais de 1.000TB de novos registros digitais de imagem advindos de Sensoriamento Remoto orbital. Tal gama de registros, que possui uma progressão geométrica crescente, é adicionada, anualmente, a incrível e extraordinária massa de dados de imagens orbitais já existentes da superfície da Terra (adquiridos desde a década de 70 do século passado). Esta quantidade maciça de registros, onde a grande maioria sequer foi processada, requer ferramentas computacionais que permitam o reconhecimento automático de padrões de imagem desejados, de modo a permitir a extração dos objetos geográficos e de alvos de interesse, de forma mais rápida e concisa. A proposta de tal reconhecimento ser realizado automaticamente por meio da integração de técnicas de Análise Espectral e de Inteligência Computacional com base no Conhecimento adquirido por especialista em imagem foi implementada na forma de um integrador com base nas técnicas de Redes Neurais Computacionais (ou Artificiais) (através do Mapa de Características Auto- Organizáveis de Kohonen SOFM) e de Lógica Difusa ou Fuzzy (através de Mamdani). Estas foram aplicadas às assinaturas espectrais de cada padrão de interesse, formadas pelos níveis de quantização ou níveis de cinza do respectivo padrão em cada uma das bandas espectrais, de forma que a classificação dos padrões irá depender, de forma indissociável, da correlação das assinaturas espectrais nas seis bandas do sensor, tal qual o trabalho dos especialistas em imagens. Foram utilizadas as bandas 1 a 5 e 7 do satélite LANDSAT-5 para a determinação de cinco classes/alvos de interesse da cobertura e ocupação terrestre em três recortes da área-teste, situados no Estado do Rio de Janeiro (Guaratiba, Mangaratiba e Magé) nesta integração, com confrontação dos resultados obtidos com aqueles derivados da interpretação da especialista em imagens, a qual foi corroborada através de verificação da verdade terrestre. Houve também a comparação dos resultados obtidos no integrador com dois sistemas computacionais comerciais (IDRISI Taiga e ENVI 4.8), no que tange a qualidade da classificação (índice Kappa) e tempo de resposta. O integrador, com classificações híbridas (supervisionadas e não supervisionadas) em sua implementação, provou ser eficaz no reconhecimento automático (não supervisionado) de padrões multiespectrais e no aprendizado destes padrões, pois para cada uma das entradas dos recortes da área-teste, menor foi o aprendizado necessário para sua classificação alcançar um acerto médio final de 87%, frente às classificações da especialista em imagem. A sua eficácia também foi comprovada frente aos sistemas computacionais testados, com índice Kappa médio de 0,86.
Only in 2011 were acquired over 1.000TB of new digital image registers arising from orbital remote sensing. This range of data, which has a geometric progression increasing, is added annually to an extraordinary and incredible mass of data from existing satellite images of Earth's surface (acquired since the 70s of last century). This massive amount of raw data requires computational tools which allow the automatic recognition of image patterns desired to allow the extraction of geographical objects and targets of interest more quickly and concisely. The proposal for such recognition to be performed automatically through Spectral Analysis and Computational Intelligence integration, based on knowledge acquired by image experts, was implemented as an integrator based on Computational Neural Networks (via Kohonens Self-Organizing Feature Maps - SOM) and Fuzzy Logic (through Mamdani) techniques. These techniques were applied to the spectral signatures pattern formed by the quantization levels or gray levels of the corresponding pattern in each spectral band of each pattern of interest, so that the pattern classification will depend, in an inseparable manner, of the spectral signatures correlation of the six bands of the sensor, like the work of image experts. Bands 1 to 5 and 7 of the Landsat-5 satellite were used for the determination of five classes / targets of interest in cover and land occupation, in three test areas located in the State of Rio de Janeiro (Guaratiba, Mangaratiba and Magé) in this integration with comparison of results with those derived from the interpretation of the imaging expert, which was corroborated by checking the ground truth. There was also a results comparison obtained with two commercial computer systems (IDRISI Taiga and ENVI 4.8) with the integrator, regarding the quality of classification (Kappa) and response time. The integrator, with hybrid classifications (supervised and unsupervised) in its implementation, proved to be effective in multispectral automatic (unsupervised) pattern recognition and in learning of these patterns, because as the input of a new test area occurs, the lower became the process of learning, which achieve a final average accuracy o f 87%, compared to the experts classifications. Its efficacy was also demonstrated compared to systems tested, with average Kappa of 0.86.

Стилі APA, Harvard, Vancouver, ISO та ін.

20

Hamilton, Erin Kinzel. "Multiscale and meta-analytic approaches to inference in clinical healthcare data." Diss., Georgia Institute of Technology, 2013. http://hdl.handle.net/1853/47600.

Повний текст джерела

Анотація:

The field of medicine is regularly faced with the challenge of utilizing information that is complicated or difficult to characterize. Physicians often must use their best judgment in reaching decisions or recommendations for treatment in the clinical setting. The goal of this thesis is to use innovative statistical tools in tackling three specific challenges of this nature from current healthcare applications. The first aim focuses on developing a novel approach to meta-analysis when combining binary data from multiple studies of paired design, particularly in cases of high heterogeneity between studies. The challenge is in properly accounting for heterogeneity when dealing with a low or moderate number of studies, and with a rarely occurring outcome. The proposed approach uses a Rasch model for translating data from multiple paired studies into a unified structure that allows for properly handling variability associated with both pair effects and study effects. Analysis is then performed using a Bayesian hierarchical structure, which accounts for heterogeneity in a direct way within the variances of the separate generating distributions for each model parameter. This approach is applied to the debated topic within the dental community of the comparative effectiveness of materials used for pit-and-fissure sealants. The second and third aims of this research both have applications in early detection of breast cancer. The interpretation of a mammogram is often difficult since signs of early disease are often minuscule, and the appearance of even normal tissue can be highly variable and complex. Physicians often have to consider many important pieces of the whole picture when trying to assess next steps. The final two aims focus on improving the interpretation of findings in mammograms to aid in early cancer detection. When dealing with high frequency and irregular data, as is seen in most medical images, the behaviors of these complex structures are often difficult or impossible to quantify by standard modeling techniques. But a commonly occurring phenomenon in high-frequency data is that of regular scaling. The second aim in this thesis is to develop and evaluate a wavelet-based scaling estimator that reduces the information in a mammogram down to an informative and low-dimensional quantification of the innate scaling behavior, optimized for use in classifying the tissue as cancerous or non-cancerous. The specific demands for this estimator are that it be robust with respect to distributional assumptions on the data, and with respect to outlier levels in the frequency domain representation of the data. The final aim in this research focuses on enhancing the visualization of microcalcifications that are too small to capture well on screening mammograms. Using scale-mixing discrete wavelet transform methods, the existing detail information contained in a very small and course image will be used to impute scaled details at finer levels. These "informed" finer details will then be used to produce an image of much higher resolution than the original, improving the visualization of the object. The goal is to also produce a confidence area for the true location of the shape's borders, allowing for more accurate feature assessment. Through the more accurate assessment of these very small shapes, physicians may be more confident in deciding next steps.

Стилі APA, Harvard, Vancouver, ISO та ін.

21

Karoui, Moussa Sofiane. "Méthodes de séparation aveugle de sources et application à la télédétection spatiale." Phd thesis, Université Paul Sabatier - Toulouse III, 2012. http://tel.archives-ouvertes.fr/tel-00790655.

Повний текст джерела

Анотація:

Cette thèse concerne la séparation aveugle de sources, qui consiste à estimer un ensemble de signaux sources inconnus à partir d'un ensemble de signaux observés qui sont des mélanges à paramètres inconnus de ces signaux sources. C'est dans ce cadre que le travail de recherche de cette thèse concerne le développement et l'utilisation de méthodes linéaires innovantes de séparation de sources pour des applications en imagerie de télédétection spatiale. Des méthodes de séparation de sources sont utilisées pour prétraiter une image multispectrale en vue d'une classification supervisée de ses pixels. Deux nouvelles méthodes hybrides non-supervisées, baptisées 2D-Corr-NLS et 2D-Corr-NMF, sont proposées pour l'extraction de cartes d'abondances à partir d'une image multispectrale contenant des pixels purs. Ces deux méthodes combinent l'analyse en composantes parcimonieuses, le clustering et les méthodes basées sur les contraintes de non-négativité. Une nouvelle méthode non-supervisée, baptisée 2D-VM, est proposée pour l'extraction de spectres à partir d'une image hyperspectrale contenant des pixels purs. Cette méthode est basée sur l'analyse en composantes parcimonieuses. Enfin, une nouvelle méthode est proposée pour l'extraction de spectres à partir d'une image hyperspectrale ne contenant pas de pixels purs, combinée avec une image multispectrale, de très haute résolution spatiale, contenant des pixels purs. Cette méthode est fondée sur la factorisation en matrices non-négatives couplée avec les moindres carrés non-négatifs. Comparées à des méthodes de la littérature, d'excellents résultats sont obtenus par les approches méthodologiques proposées.

Стилі APA, Harvard, Vancouver, ISO та ін.

22

Lu, Liang-You, and 盧亮有. "Application of FABEMD to hyper-spectral image classification." Thesis, 2013. http://ndltd.ncl.edu.tw/handle/43229087812306432050.

Повний текст джерела

Анотація:

碩士
國立中興大學
土木工程學系所
101
Hyperspectral images provide a great number of spectral information and have been broadly applied to image classifications. However, the scattered pixel problem due to atmosphere noises and incomplete classification leading unsatisfactory classification accuracy remains to be solved. A denoising process includes noise detection and deletion. This paper integrates Fast and Adaptive Bi-dimensional Emperical Mode Decomposition (FABEMD) and Minimum Noise Fraction (MNF) as a two-step denoising process to improve classification accuracy on a hyperspectral image. Regarded as low pass filter, FABEMD decomposes a hyperspectral image into several Bi-dimensional Intrinsic Mode Functions (BIMFs) and a residue image. Some of BIMF are integrated through image fusion to extracted informative images which is subsequently classified through a SVM classifier. The proposed two-step denoising process was tested on AVIRIS Indian Pines hyperspectral image and enhanced the overall accuracy up to 98.14% on the 16-classes classification. The result obtains a significant improvement in hyperspectral classification accuracy compared to the traditional and MNF-based SVMs. The proposed two-step denoising process combining FABEMD with MNF was proven to effectively eliminate a noise effect on hyperspectral images.

Стилі APA, Harvard, Vancouver, ISO та ін.

23

THAPLIYAL, ANKITA. "CONVOLUTIONAL NETWORK FEATURE HIERARCHY FOR HYPER SPECTRAL IMAGE CLASSIFICATION." Thesis, 2021. http://dspace.dtu.ac.in:8080/jspui/handle/repository/20109.

Повний текст джерела

Анотація:

Hyper spectral image classification is the recent technology that is famous among the researchers nowadays. It is simply an application of remote sensing methodology. The results of remote sensing are basically needs to get studied by the scientists properly so that they can analyze the surface area and the target information accordingly. Then next to the study of target information further conclusions can be made about that area successfully. Remote sensing is the first step to analyze the whole process in which the satellite helps to provide several images of a particular land area or vegetation portion. These images can be obtained by using active or passive remote sensing depending on the choice of user. As soon as the images are received by the sensors we not only analyse them in visible spectrum, but we do recognize them in ultra violet and infrared region of the electromagnetic spectrum. This type of technique is known as hyper spectral imaging. We use hyper spectral sensors to perform this type of imaging. This method has so many advantages over multispectral imaging in which number of spectral band information is comparatively less. Since the number of bands in hyper spectral imaging is greater than the band information in multi spectral imaging, the recognition of images and target is more specified and accurate for hyper spectral data. More significant information is obtained through hyper spectral imaging. Since we receive the data through hyper spectral imaging we need to apply the upcoming tasks to know the target area in deeper way. As soon as the input is received in the form of images that are now three dimensional due to the hyper spectral view, we need to classify these images into the categories they are having. For instance, we get the information of a vegetation area we need to classify this three dimensional image data into the different categories of vegetation in that particular portion of land. This whole process is known as image classification which is the latest topic for machine learning methods. The use of deep neural networks at present helps in doing the classification of large number of images at a time with much more accuracy and reduced complexity. In past few decades many researchers have provided their own supervised models to implement the image classification over a huge dataset to classify the images successfully. But due to the drawbacks like less accuracy and higher complexity, these models have been over take by convolutional neural networks. Supervised technology is a type of machine learning task where the model learn itself on the basis of input and the outputs provided at the time of training. The methods like SVM and CNN are supervised methods that we us for the classification purpose. Hence in this project instead of using multi spectral data, we have discussed the use of hyper spectral data. This chapter consist of six chapters. In chapter 1 we are discussing the basic of remote sensing and its types. Chapter 2 will tell us about the type of imaging method and their advantages and disadvantages so that we can prefer the suitable one to perform the objective. In chapter 3 we are looking over the multiple supervised methods like SVM, CNN and ANN that helps in the classification of hyper spectral image data. Chapter 4 and 5 are the discussion of latest model with increased accuracy and reduced complexity.

Стилі APA, Harvard, Vancouver, ISO та ін.

24

He, Yuan. "Hyper-spectral image processing using high performance reconfigurable computers." 2004. http://etd.utk.edu/2004/HeYuan.pdf.

Повний текст джерела

Анотація:

Thesis (M.S.)--University of Tennessee, Knoxville, 2004.
Title from title page screen (viewed May 17, 2004). Thesis advisor: Gregory Peterson. Document formatted into pages (x, 123 p. : col. ill.). Vita. Includes bibliographical references (p. 74-77).

Стилі APA, Harvard, Vancouver, ISO та ін.

25

Zambon, Francesca. "Spectral Analysis and Classification of Hyper- and Multispectral Data of Planetary Surfaces." Tesi di dottorato, 2014. http://www.fedoa.unina.it/9846/1/tesi_dottorato_zambon.pdf.

Повний текст джерела

Анотація:

L’obiettivo principale della Tesi di Dottorato è stato quello di utilizzare diverse tecniche di analisi per elaborare ed estrarre informazioni da dati multispettrali e iperspettrali di superfici planetarie provenienti da missioni spaziali. Sono stati analizzati i dati forniti dallo spettrometro (MASCS) e della camera (MDIS) a bordo della missione Messanger, dedicata a Mercurio e quelli dello spettrometro VIR a bordo della missione Dawn volta allo studio degli asteroidi Vesta e Cerere. Sono stati analizzati e classificati i dati di VIRS e MDIS traminte due metodi di classificazione: ISODATA e il medoto della distanza minima. E'stata poi fatta l'analisi spettrale delle unità bright di Vesta, tramite lo studio di parametri spettrali, quali centri e profondità delle bande e band area ratio al fine di determinarne la mineralogia. E' stato infine applicato a queste regioni il linear spectral unmixing per stabilire la composizione di queste unità.

Стилі APA, Harvard, Vancouver, ISO та ін.

26

Liang, Jie. "Spectral-spatial Feature Extraction for Hyperspectral Image Classification." Phd thesis, 2016. http://hdl.handle.net/1885/111995.

Повний текст джерела

Анотація:

As an emerging technology, hyperspectral imaging provides huge opportunities in both remote sensing and computer vision. The advantage of hyperspectral imaging comes from the high resolution and wide range in the electromagnetic spectral domain which reflects the intrinsic properties of object materials. By combining spatial and spectral information, it is possible to extract more comprehensive and discriminative representation for objects of interest than traditional methods, thus facilitating the basic pattern recognition tasks, such as object detection, recognition, and classification. With advanced imaging technologies gradually available for universities and industry, there is an increased demand to develop new methods which can fully explore the information embedded in hyperspectral images. In this thesis, three spectral-spatial feature extraction methods are developed for salient object detection, hyperspectral face recognition, and remote sensing image classification. Object detection is an important task for many applications based on hyperspectral imaging. While most traditional methods rely on the pixel-wise spectral response, many recent efforts have been put on extracting spectral-spatial features. In the first approach, we extend Itti's visual saliency model to the spectral domain and introduce the spectral-spatial distribution based saliency model for object detection. This procedure enables the extraction of salient spectral features in the scale space, which is related to the material property and spatial layout of objects. Traditional 2D face recognition has been studied for many years and achieved great success. Nonetheless, there is high demand to explore unrevealed information other than structures and textures in spatial domain in faces. Hyperspectral imaging meets such requirements by providing additional spectral information on objects, in completion to the traditional spatial features extracted in 2D images. In the second approach, we propose a novel 3D high-order texture pattern descriptor for hyperspectral face recognition, which effectively exploits both spatial and spectral features in hyperspectral images. Based on the local derivative pattern, our method encodes hyperspectral faces with multi-directional derivatives and binarization function in spectral-spatial space. Compared to traditional face recognition methods, our method can describe distinctive micro-patterns which integrate the spatial and spectral information of faces. Mathematical morphology operations are limited to extracting spatial feature in two-dimensional data and cannot cope with hyperspectral images due to so-called ordering problem. In the third approach, we propose a novel multi-dimensional morphology descriptor, tensor morphology profile~(TMP), for hyperspectral image classification. TMP is a general framework to extract multi-dimensional structures in high-dimensional data. The n-order morphology profile is proposed to work with the n-order tensor, which can capture the inner high order structures. By treating a hyperspectral image as a tensor, it is possible to extend the morphology to high dimensional data so that powerful morphological tools can be used to analyze hyperspectral images with fused spectral-spatial information. At last, we discuss the sampling strategy for the evaluation of spectral-spatial methods in remote sensing hyperspectral image classification. We find that traditional pixel-based random sampling strategy for spectral processing will lead to unfair or biased performance evaluation in the spectral-spatial processing context. When training and testing samples are randomly drawn from the same image, the dependence caused by overlap between them may be artificially enhanced by some spatial processing methods. It is hard to determine whether the improvement of classification accuracy is caused by incorporating spatial information into the classifier or by increasing the overlap between training and testing samples. To partially solve this problem, we propose a novel controlled random sampling strategy for spectral-spatial methods. It can significantly reduce the overlap between training and testing samples and provides more objective and accurate evaluation.

Стилі APA, Harvard, Vancouver, ISO та ін.

27

Senapati, Subhrajyoti. "Unsupervised Classification of Hyperspectral Images based on Spectral Features." Thesis, 2015. http://ethesis.nitrkl.ac.in/7177/1/Unsupervised_Senapati_2015.pdf.

Повний текст джерела

Анотація:

In this world of Big Data, large quantities of data are been created everyday from all the type of visual sensors available in the hands of mankind. One important data is that we obtain from satellite of the land image. The application of these data are numerous. They have been used in classification of land regions, change detection of an area over a period of time, detecting different anomalies in the area and so on. As data is increasing at a high rate, so manually doing these jobs is not a good idea. So, we have to apply automated algorithms to solve these jobs. The images we see generally consists of visible light in Red, Green and Blue bands, but light of different wavelength differ in their properties of passing obstacle. So, there has been considerable research going on continuous spectra images. These images are called Hyper-spectral Image. In this thesis, I have gone through many classic machine learning algorithms like K-means, Expectation Maximization, Hierarchical Clustering, some out of box methods like Unsupervised Artificial DNA Classifier, Spatial Spectral Information which integrates both features to get better classification and a variant of Maximal Margin Clustering which uses K-Nearest Neighbor algorithm to cross validate and get the best set to separate. Sometimes PCA is used get best features from the dataset. Finally all the results are compared

Стилі APA, Harvard, Vancouver, ISO та ін.

28

Du, Zheng. "Integration of Spatial and Spectral Information for Hyperspectral Image Classification." 2008. http://trace.tennessee.edu/utk_graddiss/431.

Повний текст джерела

Анотація:

Hyperspectral imaging has become a powerful tool in biomedical and agriculture fields in the recent years and the interest amongst researchers has increased immensely. Hyperspectral imaging combines conventional imaging and spectroscopy to acquire both spatial and spectral information from an object. Consequently, a hyperspectral image data contains not only spectral information of objects, but also the spatial arrangement of objects. Information captured in neighboring locations may provide useful supplementary knowledge for analysis. Therefore, this dissertation investigates the integration of information from both the spectral and spatial domains to enhance hyperspectral image classification performance. The major impediment to the combined spatial and spectral approach is that most spatial methods were only developed for single image band. Based on the traditional singleimage based local Geary measure, this dissertation successfully proposes a Multidimensional Local Spatial Autocorrelation (MLSA) for hyperspectral image data. Based on the proposed spatial measure, this research work develops a collaborative band selection strategy that combines both the spectral separability measure (divergence) and spatial homogeneity measure (MLSA) for hyperspectral band selection task. In order to calculate the divergence more efficiently, a set of recursive equations for the calculation of divergence with an additional band is derived to overcome the computational restrictions. Moreover, this dissertation proposes a collaborative classification method which integrates the spectral distance and spatial autocorrelation during the decision-making process. Therefore, this method fully utilizes the spatial-spectral relationships inherent in the data, and thus improves the classification performance. In addition, the usefulness of the proposed band selection and classification method is evaluated with four case studies. The case studies include detection and identification of tumor on poultry carcasses, fecal on apple surface, cancer on mouse skin and crop in agricultural filed using hyperspectral imagery. Through the case studies, the performances of the proposed methods are assessed. It clearly shows the necessity and efficiency of integrating spatial information for hyperspectral image processing.

Стилі APA, Harvard, Vancouver, ISO та ін.

29

Schwartzkopf, Wade Carl. "Maximum likelihood techniques for joint segmentation-classification of multi-spectral chromosome images." Thesis, 2002. http://wwwlib.umi.com/cr/utexas/fullcit?p3110690.

Повний текст джерела

Стилі APA, Harvard, Vancouver, ISO та ін.

30

Lu, Wern Horn, and 盧文鴻. "A Study of Applying Hyper-Rectangle Learning Model to Interpret the Classification of Remote Sensing Image." Thesis, 2000. http://ndltd.ncl.edu.tw/handle/95933853781797954300.

Повний текст джерела

Анотація:

碩士
中華大學
土木工程學系碩士班
88
Abstract Recently, many nations comprehensively apply remote sensing method, to serve as the important data resource of analysis and decision planning. The application of this in this aspect has also been more and more popular. The remote sensing data coverage is provided with both comprehensive and up-to-date characteristics, able to serve as a kind of effective survey tool for build environmental resource database with convenient service. This research attempts to engage in image classification with the Hyper- Rectangles Learning Model in artificial intellectual field. This model belongs to case base method. That is, through experience to attain the feedback revision objective, but starting from super space geometric concept, to store the past data in hyper rectangles structure, not only saving memory but also with systematic significance; in particular it can attain perfect accuracy in training stage which can hardly be attained in general traditional classification mode. The research, in view of the characteristics of classification, innovates the Hyper-Rectangles Learning Model, able to considerably upgrade accuracy ratio. In order to verify the classification ability of the model, theoretical math function is applied as a test example, then the SPOT satellite multispectral image applied to Tzengwen Reservoir water collection area is applied to perform land cover interpretation, with input variables in adopting the 6 classification features provided by specialists and scholars. The result shows that either five model test examples or the applied example , The accuracy of applying Hyper-Rectangles Learning Model is better than Artificial Neural Network of BPN .

Стилі APA, Harvard, Vancouver, ISO та ін.

31

Chang, Chao-Chun, and 張超群. "Mathematical Model Integrating Spectral and Spatial Information for Remotely Sensed Hyperspectral Image Classification." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/95361932830476596947.

Повний текст джерела

Анотація:

碩士
國立臺中教育大學
數學教育學系
103
Hyperspectral image data with hundreds of measured bands potentially provide more accurate and detailed spectral information for classification. However, hyperspectral image with pixel location can provide additional accurate and detailed spatial information for classification, due to the data points of each class tended to form a cluster by its spatial information. Therefore, hyperspectral remote sensing image classification integrating spectral and spatial information is a popular issue in recent years. Hence, in this study, two methods utilizing both spectral and spatial information from data on hyperspectral image classification are proposed, particularly for small sample size problems. The first one combines the concept of regularization spectral and spatial distance as two weights to design a classifier, named spectral and spatial distance classifier (SSDC). The second one is a dimensionality reduction method to extract features by introducing spectral and spatial information into the design of scatter matrices, named spectral and spatial feature extraction (SSFE). The experimental results show that the proposed classifier SSDC can achieve remarkable performance on hyperspectral image classification even in small sample size scenarios. Moreover the proposed feature extraction method SSFE can achieve better classification accuracy than some existing feature exctration algorithms for most of the classifiers. In summary, the results of this study provide more efficient ways to integrate spectral and spatial information for hyperspectral image classification.

Стилі APA, Harvard, Vancouver, ISO та ін.

32

Cheng, Ying-Ying, and 鄭盈盈. "Hyperspectral image classification via integration of joint sparse representation with spectral and spatial information." Thesis, 2019. http://ndltd.ncl.edu.tw/handle/7egfn5.

Повний текст джерела

Анотація:

碩士
國立臺灣大學
土木工程學研究所
107
Machine learning algorithms using self-learning to improve performance have become increasingly popular in the processing and analysis of remote sensing image data. The advantage of machine learning is that it is not necessary to know the prior characteristics of data in advance, and the data distribution does not have to be normally distributed, so it is more able to describe the actual distribution of remote sensing image data. Most important information of high-dimensional data (e.g. hyperspectral images) is mainly clustered in low-dimensional subspace. Moreover, pixels belonging to the same class are usually distributed in the same low-dimensional subspace. Therefore, how to reduce the dimensionality for classification has become the major issue for hyperspectral image analysis. Considerable researches have been dedicated to hyperspectral image classification via sparse representation methods over the past decade. Sparse representation has shown good performance in signal reconstruction and can be used to process data with sparse properties, so it is quite suitable for hyperspectral images analysis. On the basis of sparse representation method, the paper consists of three main parts for discussion. First, the method for dictionary construction will be introduced. With machine learning algorithm, dictionaries can be obtained by training samples of provided spectral signal. Second, the solutions for sparse coefficients optimization, such as orthogonal matching pursuit is tested for experiment analysis. Third, the model of joint sparse representation for hyperspectral image classification will be put forward. In the proposed model, the spectral and spatial information are integrated into the joint sparse representation simultaneously to improve the efficiency and accuracy of the hyperspectral image classification.

Стилі APA, Harvard, Vancouver, ISO та ін.

33

"Image classification of spatially heterogeneous land use type based on structural composition of spectral classes." Chinese University of Hong Kong, 1991. http://library.cuhk.edu.hk/record=b5886954.

Повний текст джерела

Анотація:

Chan, King-Chong.
Thesis (M.Phil.) -- Chinese University of Hong Kong, 1991.
Bibliography: leaves 150-163.
Abstract --- p.i
Acknowledgements --- p.ii
Figures --- p.vii
Tables --- p.x
Chapter Chapter 1 --- Introduction --- p.i
Chapter 1.1 --- Background --- p.2
Chapter 1.2 --- Objectives --- p.5
Chapter 1.3 --- Hypotheses --- p.6
Chapter 1.4 --- Organization of the Thesis --- p.6
Chapter Chapter 2 --- Literature Review --- p.8
Chapter 2.1 --- Land Use and Land Cover --- p.10
Chapter 2.2 --- Informational Classes and Spectra I Classes --- p.11
Chapter 2.3 --- Simple Per-Pixel Classification Method --- p.12
Chapter 2.4 --- Scene Noise and Boundary Effect --- p.14
Chapter 2.5 --- Using Filtered Data --- p.16
Chapter 2 .6 --- Textura1 Classifier --- p.18
Chapter 2.7 --- Contextual Classifier --- p.22
Chapter 2.8 --- Geographic Information System (GIS) --- p.24
Chapter 2.9 --- Expert System and Artificial Intelligence (AI) --- p.25
Chapter 2.10 --- Concluding Remarks --- p.27
Chapter Chapter 3 --- Methodology --- p.30
Chapter 3.1 --- Spectral Class Composition Method (SCCM) --- p.32
Chapter 3.1.1 --- The Concept of the Spectral Class Composition Method --- p.32
Chapter 3.1.2 --- Unsupervised Classification Process --- p.39
Chapter 3.1.3 --- Training Process --- p.39
Chapter 3.1.4 --- Proportion Counting --- p.40
Chapter 3.1.5 --- Number of Spectral Class --- p.41
Chapter 3.1.6 --- Window Size --- p.42
Chapter 3.1.7 --- Transect Process --- p.43
Chapter 3.1.8 --- Classification Task --- p.45
Chapter 3.1.9 --- Summary --- p.47
Chapter 3.2 --- Research Design --- p.49
Chapter 3.2.1 --- Study Area --- p.49
Chapter 3.2.2 --- Data and Instruments Used --- p.51
Chapter 3.2.3 --- C1assification Scheme --- p.51
Chapter 3.2.4 --- Accuracy Assessment --- p.52
Chapter Chapter 4 --- Results and Discussion I--- Examining the Relationship Between Land Use and Spectral Classes --- p.55
Chapter 4. 1 --- Unsupervised Classification --- p.57
Chapter 4.1.1 --- Unsupervised Classification Process --- p.57
Chapter 4.1.2 --- Unsupervised Classification Results --- p.58
Chapter 4.1.3 --- Difference Between Spectral Class Maps --- p.65
Chapter 4.2 --- Training Process --- p.68
Chapter 4.2.1 --- Definition of Training Process --- p.68
Chapter 4.2.2 --- Selection of Training Sites --- p.69
Chapter 4.2.3 --- Spectral Class Composition Data Extracted from the Training Sites --- p.70
Chapter 4.2.4 --- Spectral Heterogeneous Characteristics of Land Use Types --- p.73
Chapter 4.2.5 --- Different Number of Spectral Classes --- p.77
Chapter 4.2.6 --- Similar Composition Results in Some Land Use Types --- p.80
Chapter 4.2.7 --- Using Spectra1 Class Composition Data as Rules of Classification --- p.81
Chapter 4.3 --- Proportion Counting --- p.83
Chapter 4.3.1 --- Window-Based Proportion Counting Process --- p.83
Chapter 4.3.2 --- Transect Process --- p.85
Chapter 4.3.3 --- Variation of Spectra I Class Proportion within a Land Use Type --- p.91
Chapter 4.3.4 --- Variation of Spectral Class Proportion among Land Use Types --- p.95
Chapter 4.4 --- Summary --- p.103
Chapter Chapter 5 --- Resu1ts and Discussion II --- Classification and Accuracy Assessment --- p.104
Chapter 5.1 --- Rule-Based Land Use Classification --- p.106
Chapter 5.1.1 --- Derivation of Rules for Classification --- p.106
Chapter 5.1.2 --- Using Rules for Classification --- p.106
Chapter 5.1.3 --- Modification of the Rules --- p.109
Chapter 5.1.4 --- C1assification Resu11s --- p.109
Chapter 5.2 --- Accuracy Assessment --- p.118
Chapter 5.2.1 --- Accuracy Assessment Process --- p.118
Chapter 5.2.2 --- Analysis of Error Matrices --- p.123
Chapter 5.2.3 --- Comparison Between Spectral Class Composition Method and Simple Per-Pixel Method --- p.126
Chapter 5.2.4 --- Discussion on the Resui.ts of Producer's and User's Accuracy --- p.130
Chapter 5.2.5 --- Discussion on Number of Spectral Classes --- p.132
Chapter 5.2.6 --- Discuss i on on Window Size --- p.134
Chapter 5 .3 --- Summary --- p.136
Chapter Chapter 6 --- Conclusion --- p.138
Chapter 6.1 --- Summary --- p.139
Chapter 6.2 --- Limitations and Problems --- p.142
Chapter 6.3 --- Contribution --- p.147
Chapter 6.4 --- Further Research --- p.148
Bibliography --- p.150

Стилі APA, Harvard, Vancouver, ISO та ін.

34

Luu, Thi Phuong Mai. "Wetland Habitat Studies using various Classification Techniques on Multi-Spectral Landsat Imagery: Case study: Tram chim National Park, Dong Thap Vietnam." Master's thesis, 2009. http://hdl.handle.net/10362/2634.

Повний текст джерела

Анотація:

Dissertation submitted in partial fulfilment of the requirements for the Degree of Master of Science in Geospatial Technologies
Wetland is one of the most valuable ecological systems in nature. Wetland habitat is a set of comprehensive information of wetland distribution, wetland habitat types are essential to wetland management programs. Maps of wetland should provide sufficient detail, retain an appropriate scale and be useful for further mapping and inventory work (Queensland wetland framework). Remotely sensed image classification techniques are useful to detect vegetation patterns and species combination in the inaccessible regions. Automated classification procedures are conducted to save the time of the research. The purpose of the research was to develop a hierarchical classification approach that effectively integrate ancillary information into the classification process and combines ISODATA (iterative self-organizing data analysis techniques algorithm) clustering, Maximum likelihood and rule-based classifier. The main goal was to find out the best possible combination or sequence of classifiers for typically classifying wetland habitat types yields higher accuracy than the existing classified wetland map from Landsat ETM data. Three classification schemes were introduced to delineate the wetland habitat types in the idea of comparison among the methods. The results showed the low accuracy of different classification schemes revealing the fact that image classification is still on the way toward a fine proper procedure to get high accuracy result with limited effort to make the investigation on sites. Even though the motivation of the research was to apply an appropriate procedure with acceptable accuracy of classified map image, the results did not achieve a higher accuracy on knowledge-based classification method as it was expected. The possible reasons are the limitation of the image resolution, the ground truth data requirements, and the difficulties of building the rules based on the spectral characteristics of the objects which contain high mix of spectral similarities.

Стилі APA, Harvard, Vancouver, ISO та ін.

35

王欣萍. "A Study of Spectral Feature Extraction Methods for Crops Classification Base on WorldView-2 Image." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/81627368686318529008.

Повний текст джерела

Анотація:

碩士
逢甲大學
都市計畫與空間資訊學系
103
Automatically classification is the most popular method to do classifying on satellite image, and there have high ratio to use multispectral images on automatically classification in the past. Nevertheless, multispectral images has less spectral information, floristic composition has similar spectral characteristic curve is tough to be discriminated so that it just can discriminate between vegetation and non-vegetation, and can’t ascertain crop classes of farmland fields precisely and attentively, it’s still not enough to achieve the demand of agriculture management under the monitoring purpose; However, following the progressing of remote sensing technology, the bands of WorldView-2 satellite images are upgraded to eight bands providing more spectral information than past and broadening more characteristic varieties throughout the spatial characteristic extraction. Therefore, it’s a major issue that how to use high dimensional data to do an efficient pre-process to decrease dimensions and extract the classification information is benefit to crops classes of farmland fields, and does it useful to WorldView-2 satellite images classification. For ongoing research, the primary step is collecting WorldView-2 satellite images and farmland field vector data to proceed calculation and extraction of farmland field feature on satellite images, comparing the class for diversified feature variety combination on each states. The classification features in this research will be chose by the following methods: 1. spectral characteristic curve selection, 2. characteristic fusion automatically and extraction, 3. characteristic semiautomatic selection (i.e. classification features have been chosen by means of characteristic semiautomatic selection then proceed features fusion and extraction), the following step is comparing these three kinds of classification result of combination of variables, and then focus on the varieties are extracted by means of higher classification precise combination to analysis the relationship of features weight for each target crops. According to the research result to verify the method of semiautomatic characteristic selection on classification features extraction is more optimal. The entire accuracy achieve 85~94% (kappa value: 0.74~0.85), thus, before proceeding automatically feature extraction and fusion, able to bring up the crops features description of features extracting and fusing quality by getting rid of the varieties may cause classification noises throughout the spectral characteristic curve. Regarding to classification method, Support Vector Machine Classifier perform much better than Decision tree Classifer or Back Propagation Network Classifier.

Стилі APA, Harvard, Vancouver, ISO та ін.

36

Zhang, Qiang Liu Xiuwen. "Appearance-based classification and recognition using spectral histogram representations and hierarchical learning for OCA." Diss., 2005. http://etd.lib.fsu.edu/theses/available/etd-04082005-133212/.

Повний текст джерела

Анотація:

Thesis (M.S.)--Florida State University, 2005.
Advisor: Dr. Xiuwen Liu, Florida State University, College of Arts and Sciences, Dept. of Computer Science. Title and description from dissertation home page (viewed June 10, 2005). Document formatted into pages; contains x, 50 pages. Includes bibliographical references.

Стилі APA, Harvard, Vancouver, ISO та ін.

37

Chang, Chia-Hao, and 張嘉豪. "Fusion of Synthetic Aperture Radar Image and Satellite Remote Sensing Multi-spectral Images on Forest Land Cover Classification." Thesis, 2014. http://ndltd.ncl.edu.tw/handle/53083677850221470277.

Повний текст джерела

Анотація:

碩士
國立屏東科技大學
森林系所
102
Forest inventory requires considerable manpower and material resources. Therefore, using the remote sensing techniques and field surveying to reduce inventory cost and promote the efficiency is a common method in forest inventory. Remote sensing includes active radar sensor data and passive optical sensor data. These sensors provide different information, therefor integration two different type images which can be beneficial for many applications, such as environmental change research, disaster monitoring, land cover classification, and vegetation regeneration. Conventional telemetry methods for land cover classification are mainly in optical images, but are affected by weather and night that cannot to acquire image data. Synthetic Aperture Radar (SAR) relies on microwave radiation, and is not affected by the weather and night. Combining SAR and optical images to promote the forest land cover classification accuracy is main purpose of this study. In this study, we used the ALOS PALSAR L band images with a wavelength of 0.25 m and SPOT-4 multispectral satellite images. To combinating the spectrum features and roughness information, to improve classification accuracy. We use ERDAS IMAGINE 9.2 software to analyze and the process of SAR image. The result showed that Gamma filter was better than Lee filter and Frost filter. SAR images and multispectral satellite images were combined by Intensity, Hue and Saturation (IHS), Principal Component Analysis (PCA) and Wavelet Transformation (WT). We used IHS images, PCA images, WT images and SPOT-4 images to classify the forest land cover classification using Maximum Likelihood Method (MLC). The results showed that overall accuracy was 83.86% and Kappa was 0.8152 in IHS, overall accuracy was 82.44% and Kappa was 0.7989 in PCA and overall accuracy was 72.86% and Kappa was 0.6889 by WT. These results are better than those based on SPOT-4 images whose overall accuracy was 65.71% and Kappa was 0.6052. Results indicate that the fusion of SAR and optical images will improve the classification accuracy by approximately 18.15%, thereby improving forest land cover classification.

Стилі APA, Harvard, Vancouver, ISO та ін.

38

Fauvel, Mathieu. "Spectral and Spatial Methods for the Classification of Urban Remote Sensing Data." Phd thesis, 2007. http://tel.archives-ouvertes.fr/tel-00258717.

Повний текст джерела

Анотація:

Lors de ces travaux, nous nous sommes intéressés au problème de la classification supervisée d'images satellitaires de
zones urbaines. Les données traitées sont des images optiques à très hautes résolutions spatiales: données panchromatiques à très haute résolution spatiale (IKONOS, QUICKBIRD, simulations PLEIADES) et des images hyperspectrales (DAIS, ROSIS).
Deux stratégies ont été proposées.
La première stratégie consiste en une phase d'extraction de caractéristiques spatiales et spectrales suivie d'une phase de classification. Ces caractéristiques sont extraites par filtrages morphologiques : ouvertures et fermetures géodésiques et filtrages surfaciques auto-complémentaires. La classification est réalisée avec les machines à vecteurs supports (SVM)
non linéaires. Nous proposons la définition d'un noyau spatio-spectral utilisant de manière conjointe l'information spatiale
et l'information spectrale extraites lors de la première phase.\\
La seconde stratégie consiste en une phase de fusion de données pre- ou post-classification. Lors de la fusion postclassification,
divers classifieurs sont appliqués, éventuellement sur plusieurs données issues d'une même scène (image panchromat
ique, image multi-spectrale). Pour chaque pixel, l'appartenance à chaque classe est estimée à l'aide des classifieurs. Un schém
a de fusion adaptatif permettant d'utiliser l'information sur la fiabilité locale de chaque classifieur, mais aussi l'information globale disponible a priori sur les performances de chaque algorithme pour les différentes classes, est proposé
.
Les différents résultats sont fusionnés à l'aide d'opérateurs flous.
Les méthodes ont été validées sur des images réelles. Des
améliorations significatives sont obtenues par rapport aux méthodes publiées dans la litterature.

Стилі APA, Harvard, Vancouver, ISO та ін.

39

Tung, Do Thanh, and 杜青松. "Assessment of the Grey-Level Co-occurrence Matrix for Land Use/Land Cover Classification using Multi-spectral UAV Image." Thesis, 2016. http://ndltd.ncl.edu.tw/handle/9mw9xe.

Повний текст джерела

Анотація:

碩士
逢甲大學
都市計畫與空間資訊學系
104
The application of UAV has been popular in recent years due to their advantages. The UAV systems are more advantageous as compared with other manned aircraft systems. The main advantages of UAV systems are that they can fly in high risk location, unreachable areas, and at very low altitude close to the objects without threatening human life. Moreover, the UAV images also can exhibit ground surface characteristics in very high spatial resolution. Thus, the level of detail present in the UAV image has increased considerably when compared to the other multispectral satellite images and aerial photos. In this research, multispectral UAV images have been used to extract texture features for land cover/land use classification. Five cover types have been classified based on textural/spectral combination. The texture estimation is normally based on the grey-level co-occurrence matrix (GLCM) method. The texture features are extracted from UAV near infrared band by using four textural parameters (ASM, CON, ENT and VAR), fifteen window sizes (from 3x3 to 59x59) and two quantization levels (16 and 32). The supervised maximum likelihood algorithm is selected to apply to the four UAV spectral bands combined with each textural parameter independently, and to the four spectral bands combined with four textural parameters. The classification accuracy is measured by kappa coefficient calculated from confusion matrices. The main aim of this research is to evaluate the possibility of using texture features extracted by GLCM as additional information for UAV images to tackle the problem in relation to the increased internal spectral-radiometric variation of land cover types and spectral resolution limitation of UAV images. The research results show that the classification accuracy is significantly improved when the texture features are added to the UAV spectral images. The improvement of classification accuracy appeared to be different by different texture features of each cover type. The results of this research are essential for evaluating which texture feature is useful for detail and accurate land use/land cover classification.

Стилі APA, Harvard, Vancouver, ISO та ін.

40

Hung, Mi-Chi, and 洪蜜琪. "Integration of Spectral and Spatial Information in Image Classification- An Application of the Cover and Management Factor (C Value)." Thesis, 2007. http://ndltd.ncl.edu.tw/handle/79115002930716131454.

Повний текст джерела

Анотація:

碩士
中興大學
土木工程學系所
95
Located in sub-tropical and seimic areas, Taiwan frequently bears the severe threats from both typhoons and earthquakes, which could induce debris flows and landslides. The occurrence of landslide is mostly related with the land cover, so how to efficiently acquire the information of the land cover is the critical issue of disaster mitigation. Traditionally it spends a lot of manpower and time to survey the land cover by in situ investigation. In order to get the classifications of land cover rapidly and timely, this research integrates Maximum Likelihood Method and Watershed Segmentation Algorithm into a supervised classifier. The original image is computed into brightness gradient for the segmentation of the image. The proposed method can solve the misjudgment of isolated points and mixpixel of boundary in the original image. Integration of Spectral and Spatial Classification (SSC) increases the accuracy of classification. To verify the utility of SSC classifier, various images including homogeneous and non-homogeneous pictures, a public Multi-Spectral image (Purdue University), and a QuickBird satellite images (National Chung Hsing University) were tested. The result indicated that this method can decrease the misjudgment of isolated points and mixpixel on both photos and satellite images. The accuracies of SSC and ECHO (Extraction and Classification of Homogeneous Objects) of Landgrebe (1998) are 75.2% and 76.3%, respectively for the classification of purdue University. Applying this method to the watershed of Shihman Reservoir by corresponding the coverage and management factor (C Value) through the land cover classification can provide the reference of land development and the amount of soil loss.

Стилі APA, Harvard, Vancouver, ISO та ін.

41

"Clinically Relevant Classification and Retrieval of Diabetic Retinopathy Images." Master's thesis, 2012. http://hdl.handle.net/2286/R.I.15144.

Повний текст джерела

Анотація:

abstract: Diabetic retinopathy (DR) is a common cause of blindness occurring due to prolonged presence of diabetes. The risk of developing DR or having the disease progress is increasing over time. Despite advances in diabetes care over the years, DR remains a vision-threatening complication and one of the leading causes of blindness among American adults. Recent studies have shown that diagnosis based on digital retinal imaging has potential benefits over traditional face-to-face evaluation. Yet there is a dearth of computer-based systems that can match the level of performance achieved by ophthalmologists. This thesis takes a fresh perspective in developing a computer-based system aimed at improving diagnosis of DR images. These images are categorized into three classes according to their severity level. The proposed approach explores effective methods to classify new images and retrieve clinically-relevant images from a database with prior diagnosis information associated with them. Retrieval provides a novel way to utilize the vast knowledge in the archives of previously-diagnosed DR images and thereby improve a clinician's performance while classification can safely reduce the burden on DR screening programs and possibly achieve higher detection accuracy than human experts. To solve the three-class retrieval and classification problem, the approach uses a multi-class multiple-instance medical image retrieval framework that makes use of spectrally tuned color correlogram and steerable Gaussian filter response features. The results show better retrieval and classification performances than prior-art methods and are also observed to be of clinical and visual relevance.
Dissertation/Thesis
M.S. Electrical Engineering 2012

Стилі APA, Harvard, Vancouver, ISO та ін.

42

Chan, Ching-Yi, and 詹靜怡. "Using Satellite Remote Sensing Multi-Spectral Image for Classification of Forest Fire Severity and Vegetation Recovery in Huanshan Area of Taichung." Thesis, 2012. http://ndltd.ncl.edu.tw/handle/69154386209916747059.

Повний текст джерела

Анотація:

碩士
國立屏東科技大學
森林系所
100
Monitoring the ecological change via post fire after forest fire is an important issue in forest conservation. The spatial model of forest fire severity had been developed by the establishment of the relational model of burned- site land cover and satellite vegetation index. The red and near-infrared band multi-spectral satellite images, which were used to estimate the vegetation index and to discriminate analysis to develop forest fire severity, showed a higher discriminative rate. This study used the Landsat 7 ETM + infrared band (band 7,2.08 -2.35 μm) to classify and map the severity of forest fire. We found Normalized Difference Burn Ratio (dNBR) and Relativized Normalized Difference Burn Ratio (RdNBR) are effective to assess fire severity. Using object-based image analysis, which is a multilevel segmentation and a classification, is easier to evaluate and distinguish at least 3 kinds of forest fire severities. The evaluation of vegetation recovery after forest fire showed VRR value was up to 50% in the high-severity burned areas after 3 to 5 years, and VRR value was more than 95% after 10 years. NASA provides free Landsat satellite images currently. The spectral resolution and spatial resolution are suitable to assess forest fire severity and monitor the vegetation recovery after forest fire.

Стилі APA, Harvard, Vancouver, ISO та ін.

43

Omran, Mahamed G. H. "Particle swarm optimization methods for pattern recognition and image processing." Thesis, 2005. http://hdl.handle.net/2263/29826.

Повний текст джерела

Анотація:

Pattern recognition has as its objective to classify objects into different categories and classes. It is a fundamental component of artificial intelligence and computer vision. This thesis investigates the application of an efficient optimization method, known as Particle Swarm Optimization (PSO), to the field of pattern recognition and image processing. First a clustering method that is based on PSO is proposed. The application of the proposed clustering algorithm to the problem of unsupervised classification and segmentation of images is investigated. A new automatic image generation tool tailored specifically for the verification and comparison of various unsupervised image classification algorithms is then developed. A dynamic clustering algorithm which automatically determines the "optimum" number of clusters and simultaneously clusters the data set with minimal user interference is then developed. Finally, PSO-based approaches are proposed to tackle the color image quantization and spectral unmixing problems. In all the proposed approaches, the influence of PSO parameters on the performance of the proposed algorithms is evaluated.
Thesis (PhD)--University of Pretoria, 2006.
Computer Science
unrestricted

Стилі APA, Harvard, Vancouver, ISO та ін.

44

Χριστούλας, Γεώργιος. "Ανάπτυξη συστήματος επεξεργασίας δεδομένων τηλεπισκόπησης για αυτόματη ανίχνευση και ταξινόμηση περιοχών με περιβαλλοντικές αλλοιώσεις". Thesis, 2012. http://hdl.handle.net/10889/5312.

Повний текст джерела

Анотація:

Η παρούσα διατριβή είχε σαν κύριο στόχο την ανάλυση και επεξεργασία των δεδομένων SAR υπό το πρίσμα του περιεχομένου υφής για την ανίχνευση περιοχών με περιβαλλοντικές αλλοιώσεις όπως είναι οι παράνομες εναποθέσεις απορριμμάτων. Τα δεδομένα που χρησιμοποιήθηκαν προέρχονταν από τον δορυφόρο ENVISAT και το όργανο ASAR του Ευρωπαϊκού Οργανισμού Διαστήματος με διακριτική ικανότητα 12.5m και 30m για τις λειτουργίες μονής και διπλής πολικότητας αντίστοιχα καθώς και από τον δορυφόρο Terra-SAR με διακριτική ικανότητα 3m και HH πολικότητα. Χρησιμοποιήθηκαν κλασσικές τεχνικές ανάλυσης και ταξινόμησης υφής όπως GLCM, Markov Random Fields, Gabor Filters και Neural Networks. Η μελέτη προσανατολίστηκε στην ανάπτυξη νέων μεθόδων ταξινόμησης υφής για αυξημένη αποτελεσματικότητα. Χρησιμοποιήθηκαν δεδομένα πολυφασματικά και SAR. Για τα πολυφασματικά δεδομένα προτάθηκε η χρήση της spectral co-occurrence ως χαρακτηριστικό υφής που χρησιμοποιεί πληροφορία φασματικού περιεχομένου. Για τα δεδομένα SAR αναπτύχθηκε μία νέα μέθοδος ταξινόμησης η οποία βασίζεται σε συνήθεις περιγραφείς υφής (GLCM, Gabor, MRF) οι οποίοι μελετώνται για την ικανότητά τους να διαχωρίζουν ζεύγη μεταξύ τάξεων. Για κάθε ζεύγος τάξεων προκύπτουν χαρακτηριστικά υφής που βασίζονται στις στατιστικές ιδιότητες της cumulative καθώς και της πρώτης και δεύτερης τάξης αυτής. Η μέθοδος leave one out χρησιμοποιείται για τον εντοπισμό των χαρακτηριστικών που μπορούν να διαχωρίσουν τα δείγματα ανά ζεύγη τάξεων στα οποία αντιστοιχίζεται και ένας ξεχωριστός και ανεξάρτητος γραμμικός ταξινομητής. Η τελική ταξινόμηση γίνεται με τη μέθοδο της πλειοψηφίας η οποία εφαρμόζεται στο πρόβλημα των δύο τάξεων και τριών τάξεων αλλά επεκτείνεται και στο πρόβλημα των N-τάξεων δεδομένης της ύπαρξης κατάλληλων χαρακτηριστικών.
Texture characteristics of MERIS data based on the Gray-Level Co-occurrence Matrices (GLCM) are explored as far as their classification capabilities are concerned. Classification is employed in order to reveal four different land cover types, namely: water, forest, field and urban areas. The classification performance for each cover type is studied separately on each spectral band, while the combined performance of the most promising spectral bands is explored. In addition to GLCM, spectral co-occurrence matrices (SCM) formed by measuring the transition from band-to-band are employed for improving classification results. Conventional classifiers and voting techniques are used for the classification stage. Furthermore, the properties of texture characteristics are explored on various types of grayscale or RGB representations of the multispectral data, obtained by means of principal components analysis (PCA), non-negative matrix factorization (NMF) and information theory. Finally, the accuracy of the proposed classification approach is compared with that of the minimum distance classifier. A simple and effective classification method is furthermore proposed for remote sensed data that is based on a majority voting schema. We propose a feature selection procedure for exhaustive search of occurrence measures resulting from fundamental textural descriptors such as Co-occurrence matrices, Gabor filters and Markov Random Fields. In the proposed method occurrence measures, that are named texture densities, are reduced to the local cumulative function of the texture representation and only those that can linearly separate pairs of classes are used in the classification stage, thus ensuring high classification accuracy and reliability. Experiments performed on SAR data of high resolution and on a Brodatz texture database have given more than 90% classification accuracy with reliability above 95%.

Стилі APA, Harvard, Vancouver, ISO та ін.

45

Strothmann, Wolfram. "Multi-wavelength laser line profile sensing for agricultural applications." Doctoral thesis, 2016. https://repositorium.ub.uni-osnabrueck.de/handle/urn:nbn:de:gbv:700-2016110315110.

Повний текст джерела

Анотація:

This dissertation elaborates on the novel sensing approach of multi-wavelength laser line profiling (MWLP). It is a novel sensor concept that expands on the well-known and broadly adopted laser line profile sensing concept for triangulation-based range imaging. Thereby, the MWLP concept does not just use one line laser but multiple line lasers at different wavelengths scanned by a single monochrome imager. Moreover, it collects not only the 3D distance values but also reflection intensity and backscattering of the laser lines are evaluated. The system collects spectrally selective image-based data in an active manner. Thus, it can be geared toward an application-specific wavelength configuration by mounting a set of lasers of the required wavelengths. Consequently, with this system image-based 3D range data can be collected along with reflection intensity and backscattering data at multiple, selectable wavelengths using just a single monochrome image sensor. Starting from a basic draft of the idea, the approach was realized in terms of hardware and software design and implementation. The approach was shown to be feasible and the prototype performed well as compared with other state-of-the-art sensor systems. The sensor raw data can be visualized and accessed as overlayed distance images, point clouds or mesh. Further, for selected example applications it was demonstrated that the sensor data gathered by the system can serve as descriptive input for real world agricultural classification problems. The sensor data was classified in a pixel-based manner. This allows very flexible, quick and easy adaptation of the classification toward new field situations.

Стилі APA, Harvard, Vancouver, ISO та ін.

Дисертації з теми "HYPER SPECTRAL IMAGE CLASSIFICATION"

Оформте джерело за APA, MLA, Chicago, Harvard та іншими стилями