Dissertations / Theses on the topic 'IMAGE DEHAZING'
To see the other types of publications on this topic, follow the link: IMAGE DEHAZING.
Create a spot-on reference in APA, MLA, Chicago, Harvard, and other styles
Consult the top 31 dissertations / theses for your research on the topic 'IMAGE DEHAZING.'
Next to every source in the list of references, there is an 'Add to bibliography' button. Press on it, and we will generate automatically the bibliographic reference to the chosen work in the citation style you need: APA, MLA, Harvard, Chicago, Vancouver, etc.
You can also download the full text of the academic publication as pdf and read online its abstract whenever available in the metadata.
Browse dissertations / theses on a wide variety of disciplines and organise your bibliography correctly.
1
Pérez, Soler Javier. "Visibility in underwater robotics: Benchmarking and single image dehazing." Doctoral thesis, Universitat Jaume I, 2017. http://hdl.handle.net/10803/432778.
Full textAbstract:
Dealing with underwater visibility is one of the most important challenges in autonomous underwater robotics. The light transmission in the water medium degrades images making the interpretation of the scene difficult and consequently compromising the whole intervention. This thesis contributes by analysing the impact of the underwater image degradation in commonly used vision algorithms through benchmarking. An online framework for underwater research that makes possible to analyse results under different conditions is presented. Finally, motivated by the results of experimentation with the developed framework, a deep learning solution is proposed capable of dehazing a degraded image in real time restoring the original colors of the image.Una de las dificultades más grandes de la robótica autónoma submarina es lidiar con la falta de visibilidad en imágenes submarinas. La transmisión de la luz en el agua degrada las imágenes dificultando el reconocimiento de objetos y en consecuencia la intervención. Ésta tesis se centra en el análisis del impacto de la degradación de las imágenes submarinas en algoritmos de visión a través de benchmarking, desarrollando un entorno de trabajo en la nube que permite analizar los resultados bajo diferentes condiciones. Teniendo en cuenta los resultados obtenidos con este entorno, se proponen métodos basados en técnicas de aprendizaje profundo para mitigar el impacto de la degradación de las imágenes en tiempo real introduciendo un paso previo que permita recuperar los colores originales.
2
Karlsson, Jonas. "FPGA-Accelerated Dehazing by Visible and Near-infrared Image Fusion." Thesis, Mälardalens högskola, Akademin för innovation, design och teknik, 2015. http://urn.kb.se/resolve?urn=urn:nbn:se:mdh:diva-28322.
Full textAbstract:
Fog and haze can have a dramatic impact on vision systems for land and sea vehicles. The impact of such conditions on infrared images is not as severe as for standard images. By fusing images from two cameras, one ordinary and one near-infrared camera, a complete dehazing system with colour preservation can be achieved. Applying several different algorithms to an image set and evaluating the results, the most suitable image fusion algoritm has been identified. Using an FPGA, a programmable integrated circuit, a crucial part of the algorithm has been implemented. It is capable of producing processed images 30 times faster than a laptop computer. This implementation lays the foundation of a real-time dehazing system and provides a significant part of the full solution. The results show that such a system can be accomplished with an FPGA.
3
Hultberg, Johanna. "Dehazing of Satellite Images." Thesis, Linköpings universitet, Datorseende, 2018. http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-148044.
Full textAbstract:
The aim of this work is to find a method for removing haze from satellite imagery. This is done by taking two algorithms developed for images taken from the sur- face of the earth and adapting them for satellite images. The two algorithms are Single Image Haze Removal Using Dark Channel Prior by He et al. and Color Im- age Dehazing Using the Near-Infrared by Schaul et al. Both algorithms, altered to fit satellite images, plus the combination are applied on four sets of satellite images. The results are compared with each other and the unaltered images. The evaluation is both qualitative, i.e. looking at the images, and quantitative using three properties: colorfulness, contrast and saturated pixels. Both the qualitative and the quantitative evaluation determined that using only the altered version of Dark Channel Prior gives the result with the least amount of haze and whose colors look most like reality.
4
Han, Che, and 蘇哲漢. "Nighttime Image Dehazing." Thesis, 2014. http://ndltd.ncl.edu.tw/handle/3d34wx.
Full textAbstract:
碩士國立中山大學
資訊工程學系研究所
102
Image surveillance is the major means of security monitoring. Image sequences obtained through surveillance cameras are vital sources for tracking criminal incidents and causes of accident, happening mostly at night due to lacking of light and obscurity of vision. The quality of the image plays a pivotal role in providing evidence and uncovering the truth. However, almost all image processing techniques focus on daylight environment, seldom on compensating artifacts rooted from artificial light source at night or light diffusion. The low-lighting environment and color obscurity often invalidate further identification from the surveillance video acquired. The processing of images acquired at night cannot follow the paradigm of the daylight image processing. Take image dehazing for example, the removal of haze depends on the derivation of scene depth. Dark Channel Prior (DCP), using dark channel as a prior assumption, is often applied to derive scene depth from a single image. The farthest area, with the highest intensity of light, in an image corresponds to the major source of lighting – daylight, while the area closer with lower degree of light intensity, Therefore, the depth within the scene links with the amount of background light. The above observation does not hold at night. The source of light does not come from sun, rather artificial light source, e.g., street lamp or automobile headlight. The farthest area, often dark-pitch due to lack of any light source, does not have the highest light intensity. To the best of our knowledge, no research has been reported regarding the nighttime image dehazing and enhancement. In light of the demands of higher nighttime image quality, this paper proposes an image dehazing technique, incorporating the light diffusion model, artificial light source, and segmentation of moving objects within the image sequence, to restore the nighttime scene back to the daytime one. The paper, employing the dehazing and image enhancement to remove the light diffusion in a nighttime image, is composed of daytime background dehazing and nighttime image enhancement. The scene depth is derived by applying DCP to the daytime background image, producing the corresponding depth map. The haze within the scene is removed by the dehazing algorithm to restore the daytime background. The reflectance of objects in the background can be further derived by taking the daylight intensity into consideration. The position and overall intensity of the artificial light sources can be determined through the nighttime background image first. The moving objects are then segmented from the image sequence. The reflectance of moving objects can be evaluated, given the depth map obtained from the daytime image, and position and overall intensity of the artificial light sources from the nighttime counterpart. Once the reflectance of moving objects are determined, the background and moving objects can be fused together given proper daytime lighting.
5
Jyun-GuoWang and 王峻國. "Image Dehazing Using Machine Learning Methods." Thesis, 2016. http://ndltd.ncl.edu.tw/handle/9nq7c6.
Full textAbstract:
博士國立成功大學
電腦與通信工程研究所
104
In recent years, the image dehazing issue has been widely discussed. During photography in an outdoor environment, the medium in the air causes light attenuation and reduce image quality; these impacts are especially obvious in a hazy environment. Reduction of image quality results in the loss of information, which hinders image recognition systems to identify objects in the image. Removal of haze can provide a reference for subsequent image processing for specific requirements. Notably, image dehazing technology is used to maintain image quality during preprocessing. This dissertation presents machine learning methods for image haze removal and consists of two major parts. In the first part, a fuzzy inference system (FIS) model is presented. Users of this model can customize designs to generate applicable fuzzy rules from expert knowledge or data. The number of fuzzy rules is fixed. In addition, the FIS model requires substantial amounts of data and expertise; even if the model is used to develop a fuzzy system, the image output of that system may suffer from a loss of accuracy. Therefore, in the second part of this dissertation, a recurrent fuzzy cerebellar model articulation controller (RFCMAC) model with a self-evolving structure and online learning is presented to improve the FIS model. The recurrent structure in an RFCMAC is formed with internal loops and internal feedback by feeding the rule firing strength of each rule to other rules and to itself. A Takagi-Sugeno-Kang (TSK) type is used in the consequent part of the RFCMAC. The online learning algorithm consists of structure and parameter learning. The structure learning depends on an entropy measure to determine the number of fuzzy rules. The parameter learning, based on back-propagation, can adjust the shape of the membership function and the corresponding weights of the consequent part. This dissertation describes, the proposed machine learning methods and its related algorithm, applies them to various image dehazing problems, and analyzes the results to demonstrate the effectiveness of the proposed methods.
6
Chung, Yun-Xin, and 鍾昀芯. "A Study in Image Dehazing Approaches." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/j2syw6.
Full textAbstract:
碩士國立中興大學
土木工程學系所
106
Optical image is easily affected by poor weather such as rain, snow, fog and haze, which may deteriorate the quality of input image. The process of enhancing an image by eliminating smog is called defogging. The purpose of this paper is to apply dark channel prior and histogram equalization methods to remove smog from images. The dark channel prior method is improved based on the distribution of fog to promote the seed calculation for huge remotely sensed imagery. Finally, we applied the image quality assessment index to evaluate and analyze the image dehazing results. The experimental results show that the method removes smog from the foggy image and enhances the haze by maintaining contrast.
7
Chen, Ying-Ching, and 陳英璟. "Underwater image enhancement: Using WavelengthCompensation and Image Dehazing (WCID)." Thesis, 2011. http://ndltd.ncl.edu.tw/handle/94271506864231404657.
Full textAbstract:
碩士國立中山大學
資訊工程學系研究所
99
Light scattering and color shift are two major sources of distortion for underwater photography. Light scattering is caused by light incident on objects reflected and deflected multiple times by particles present in the water before reaching the camera. This in turn lowers the visibility and contrast of the image captured. Color shift corresponds to the varying degrees of attenuation encountered by light traveling in the water with different wavelengths, rendering ambient underwater environments dominated by bluish tone. This paper proposes a novel approach to enhance underwater images by a dehazing algorithm with wavelength compensation. Once the depth map, i.e., distances between the objects and the camera, is estimated by dark channel prior, the light intensities of foreground and background are compared to determine whether an artificial light source is employed during image capturing process. After compensating the effect of artifical light, the haze phenomenon from light scattering is removed by the dehazing algorithm. Next, estimation of the image scene depth according to the residual energy ratios of different wavelengths in the background is performed. Based on the amount of attenuation corresponding to each light wavelength, color shift compensation is conducted to restore color balance. A Super-Rsolution image can offer more details that must be important and necessary in low resolution underwater image. In this paper combine Gradient-Base Super Resolution and Iterative Back-Projection (IBP) to propose Cocktail Super Resolution algorithm, with the bilateral filter to remove the chessboard effect and ringing effect along image edges, and improve the image quality. The underwater videos with diversified resolution downloaded from the Youtube website are processed by employing WCID, histogram equalization, and a traditional dehazing algorithm, respectively. Test results demonstrate that videos with significantly enhanced visibility and superior color fidelity are obtained by the WCID proposed.
8
Liao, Jyun-Jia, and 廖俊嘉. "A New Transmission Map for Image Dehazing." Thesis, 2014. http://ndltd.ncl.edu.tw/handle/sz947w.
Full textAbstract:
碩士國立臺北科技大學
電腦與通訊研究所
102
Visibility of the captured outdoor images in inclement weathers, such as haze, fog and mist, usually is degraded due to the effect of absorption and scattering caused by the atmospheric particles. Such images may significantly contaminate the performance qualities of the intelligent transportation systems relying on visual feature extraction, such as traffic status detection, traffic sign recognition, vehicular traffic tracking, and so on. Recently, haze removal techniques taken in these particular applications have caught increasing attention in improving the visibility of hazy images in order to make the performances of the intelligent transportation systems more reliable and efficient. However, estimating haze from a single haze image with an actual scene is difficult for visibility restoration methods to accomplish. In order to solve this problem, we propose a haze removal method which requires a combination of two main modules：the haze thickness estimation module and the visibility restoration module. The haze thickness estimation module is based on bi-gamma modification to effectively estimate haze for transmission map. Subsequently, the visibility restoration module utilizes the transmission map to achieve the haze removal. The experimental results demonstrate that the proposed haze removal method can restore the visibility in single haze images more effectively than can other state-of-the-art methods.
9
Jui-ChiangWen and 溫瑞強. "Single image dehazing based on vector quantization." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/69190335931091186410.
Full textAbstract:
碩士國立成功大學
電腦與通信工程研究所
103
The proposed method is based on McCartney’s optical haze model and uses a novel approach to estimate transmission. According to the literature, the major problem is estimating the transmission in the model-based method. This study trains plenty of haze-free and hazy images as codebooks with LBG algorithm. Then it is used to estimate transmission with matching. In order to speed up the process, the input image is down-sampled before refining with guided image filter. It not only can reduce processing time but also can preserve the quality of restored images. RGB, dark channel, and contrast values are regarded as features while training codebooks and estimating transmission. The transmission can be selected accurately because dark channel and contrast feature have complementarity. The experiment results show that the haze-free high-intensity objects can avoid over dehazing and keep the foreground of restored images more natural. The details of recovered images are also clearer.
10
Huang, Ren-Jun, and 黃任駿. "Single Image Dehazing Algorithm with Two-objective Optimization." Thesis, 2016. http://ndltd.ncl.edu.tw/handle/33776514616653182214.
Full textAbstract:
碩士朝陽科技大學
資訊工程系
104
To record images with a camera under different climate condition, the quality of the image will be affected by the weather such as smoke, haze, rain and snow. Among them, haze is frequently an atmospheric phenomenon where dust, smoke and other dry particles obscure the clarity of the sky. In order to improve the poor quality of image due to low visibility by haze, researchers have proposed various methods to remove haze. One of them, a Proposed Dehazing Algorithm (PDA) developed by Hsieh has a good dehazing performance. However, the resulted image after dehazing changes the mood and has a poor visual sense under certain circumstance. Most dehazing performance measures are based on subjective visual to assess the pros and cons up until now. To overcome this drawback, we proposed a Proposed Optimization Dehazing Algorithm (PODA) with two-objectives evaluation, to improve the image with a good dehazing performance and maintain the mood retention. In addition, we proposed an evaluating method for dehazing image to analyze the performance of dehazing image. The developed PODA has compared with other dehazing methods using various examples. Simulation results indicate that the PODA outperforms these competing methods.
11
Lee, Yen-Ling, and 李彥玲. "The Study of Single Remotely Sensed Image Dehazing." Thesis, 2011. http://ndltd.ncl.edu.tw/handle/62207648140186276401.
Full textAbstract:
碩士國立高雄應用科技大學
土木工程與防災科技研究所
99
With the advance of the modern spatial information technology, the remotely sensed visible imagery has been used in various filed. The main characteristics of visible imagery are high spatial resolution, stable geometry, and rich information appropriate for human vision. It is also the most important reference for the disaster prevention and management applications. But it is often affected by climate. Among them, especially, fog and haze is the major influence factor when fetching the imagery information. Therefore, effective image dehazing could promote the feasibility of imagery, and decrease the weather condition prerequisite for aerial photography. There are a lot of methods for image dehazing. In this study, we mainly discuss the Retinex and Dark Channel Prior algorithms which are the most innovative technology in image dehazing. We improved the color correction problems of MSR-based Retinex method corresponding to the human vision. The dark channel prior is also improved based on distribution of fog to promote the calculation speed for huge remotely sensed imagery. This study used various imageries for experiments, and utilized the white balance as an image post-processing tool to make the color of dehazed imagery being appropriate for human visualization. Finally, we applied the image quality assessment index fo objectively evaluate and analysis the image dehazing results.
12
Wu, Yu-Chen, and 吳宥蓁. "Stereo Image Dehazing Based on Cross Bilateral Filtering." Thesis, 2017. http://ndltd.ncl.edu.tw/handle/v343vr.
Full textAbstract:
碩士國立交通大學
多媒體工程研究所
105
In this thesis, we present a novel dehazing approach for stereo images based on cross bilateral filtering. Numerous dehazing algorithms have been proposed before. Nevertheless, most of the dehazing algorithms are proposed for a single image. This will produce inconsistent results if dehazing stereo images iteratively. We simultaneously estimate scene depth and dehaze the stereo images. The proposed approach is based on the observation of depth cues in the stereo images. The main idea of using depth cues is to avoid inconsistent results and of using the cross bilateral filter is to preserve shape details. The results show that the proposed approach can get better results compared with the previous methods.
13
Tzu-YuHuang and 黃子育. "Single image dehazing using cycle consistent adversarial networks." Thesis, 2019. http://ndltd.ncl.edu.tw/handle/2cu68b.
Full textAbstract:
碩士國立成功大學
電腦與通信工程研究所
107
Computer vision is widely used in many fields, such as surveillance system, image recognition, etc. However, image quality highly affects the result and accuracy in image recognition. Images that taken at outdoor are highly affected by particles in air. Particles cause light scattering, so the images become hazy. Hazy images will highly decrease the image quality. It is not conducive to image recognition and it will decrease the accuracy. In recent years, many researchers proposed many single image dehazing methods aim to solve the problem. In this Thesis, a method based on cycle consistent adversarial networks is proposed. After adding the dark channel, depth map and other losses into the networks, our experiment has a better result both in visual and quantitative metrics.
14
Huang, Yi-Chi, and 黃義熾. "Color Transferred Convolutional Neural Networks for Image Dehazing." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/48x28e.
Full textAbstract:
碩士元智大學
資訊工程學系
107
Image dehazing is a crucial image processing step for outdoor vision systems. However, image recovered by conventional image dehazing methods that use either haze-relevant priors or heuristic cues to estimate the transmission map may not lead to an accurate enough haze removal from a single image. The most commonly observed effects are darkened and brightened artifacts on some areas of recovered images, which cause significant loss of fidelity, brightness, and sharpness in images. This thesis develops a new image dehazing method based on a color transferred image dehazing model that improves on conventional image dehazing methods. By creating the color transferred image dehazing model for removing the haze obscuration and learning the coefficients of the model using the devised convolutional neural network based deep framework as a supervised learning strategy, an image's fidelity, brightness, and sharpness can be effectively restored. The experimental results verify that the proposed method obviously outperforms the existing single image dehazing methods via both quantitative and qualitative evaluations on either synthesized or real haze images.
15
Lee, Tzu-Yen, and 李慈晏. "Image Dehazing and Rain Removal Using Digital Image and Video Processing." Thesis, 2012. http://ndltd.ncl.edu.tw/handle/06715898331651933942.
Full textAbstract:
碩士國立臺灣大學
電信工程學研究所
100
Poor visibility in bad weather is a major problem for many applications of computer vision such as surveillance, intelligent vehicles, and outdoor object recognition, etc…. The reason is that the substantial presence of atmospheric particles has significant size and distribution in the participating medium. Based on this, weather conditions can be characterized as static and dynamic cases. Specifically, static bad weather such as fog and haze caused by microscopic particles are usually spatially and temporally consistent. Oppositely, dynamic bad weather has large particles such as raindrops and snowflakes. Because spatially and temporally neighboring areas are affected by rain and snow differently, the analysis is more difficult. Under these conditions, the human viewer would be annoyed and confused. They also degrade the effectiveness of any computer vision algorithm based on small features. Therefore, it is necessary to model the visual effects for the various cases and then remove them. In this thesis, we introduce three existing typical single image dehazing methods: contrast-based [1], independent component analysis [2], and dark channel prior-based [3]. To improve the dehazing quality, we propose a robust and effective dehazing method. Unlike other existing methods, our method gives satisfactory dehazing quality during daytime and nighttime. Besides, three existing typical rain removal methods: streak-based detection [4], image-based blurring [5], and frequency-based analysis [6] are also introduced in the literature. In this follows, we design a simple but effective rain removal method by combining the average-based rain detection and block-based rain removing procedures.
16
CHEN, YI-SIANG, and 陳奕翔. "Investigations on Image Dehazing in Multi-core computing environments." Thesis, 2017. http://ndltd.ncl.edu.tw/handle/8f2t89.
Full textAbstract:
碩士靜宜大學
資訊工程學系
105
With the rapid development of drones, clear images taken by drone are crucial to assist navigation and to avoid crash. However, hazed image may be captured in harsh weather condition, such as fogs in the atmosphere or suspended solids in under water environment. To clarify the haze images, the image dehazing algorithm is used to eliminate fogs for recovering the real scene. With the algorithm of Dark Channel Prior and Guided Filter, the image dehazing can be achieved quickly. In this thesis, we implement parallel image dehazing with OpenMP developed from existing fast sequential algorithm. The main purpose of this research is to analyze the correlation between the parallel haze removal code rewritten for shared memory multi-core servers and its performances. In preliminary studies, our experimental results of haze removal parallel code with OpenMP obtained better performance.
17
SINGH, PRABH PREET. "DEHAZING USING LIFTING WAVELET AND LAPLACIAN MATTE." Thesis, 2016. http://dspace.dtu.ac.in:8080/jspui/handle/repository/15391.
Full textAbstract:
This work presents single image dehazing based on dark channel prior. Firstly, lifting
haar wavelet has been used to decompose the hazy image into approximation and
details. Then only the approximation component which is just one-fourth of the actual
image dimension is further processed for dehazing. The dark channel prior used in the
proposed work for dehazing is based on statistics of haze free images. The property
that the intensity of dark channel gives approximate thickness of the haze is used to
estimate the transmission and atmospheric light. Instead of constant airlight, proposed
method employs scene depth to estimate spatially varying atmospheric light as it truly
occurs in nature. Haze imaging model together with soft matting method has been
used in this work to obtain a high quality haze free image. Experimental results
demonstrate that the proposed approach produces better results as color fidelity and
contrast of haze free image are improved and no over saturation in the sky region is
observed. Further, with the use of lifting wavelet transform, reduction in
computational time by a factor of two to three as compared to the conventional
approach has been observed.
18
Tsai, Yu-Tai, and 蔡雨泰. "Single Image Dehazing, Rain/Snow Removal and Underwater Enhancement Using Digital Image Processing." Thesis, 2014. http://ndltd.ncl.edu.tw/handle/20896594117992561651.
Full textAbstract:
碩士國立臺灣大學
電信工程學研究所
102
Poor visibility in bad environment is a major problem for many applications of computer vision such as surveillance, intelligent vehicles, and outdoor object recognition, etc…. The reason is that the substantial presence of atmospheric particles has significant size and distribution in the participating medium. Based on this, weather conditions can be characterized as steady and dynamic cases. Specifically, steady bad weather such as fog and haze caused by microscopic particles is usually spatially and temporally consistent. Oppositely, dynamic bad weather such as rain and snow in made up of large particles. Because spatially and temporally neighboring areas are affected by rain and snow differently, the analysis is more difficult. However, the poor visibility in underwater photography is caused by light scattering and color shift. Color shift corresponds to the varying degrees of attenuation encountered by light traveling in the water with different wavelengths, rendering ambient underwater environments dominated by bluish tone. Light scattering is caused by light incident on objects reflected and deflected multiple times by particles present in the water before reaching the camera. This in turn lowers the visibility and contrast of the image captured. Under these conditions, the human viewer would be annoyed. They also degrade the effectiveness of any computer vision algorithm based on small features and varying degrees of attenuation. Therefore, it is necessary to model the visual effects for the various cases and then remove them. In this thesis, we introduce three existing typical single image dehazing methods: contrast-based, independent component analysis, and dark channel prior-based. To improve the dehazing quality, we propose a robust and effective dehazing method. Unlike other existing methods, there is the satisfactory dehazing quality during daytime and nighttime by our methods. And then, four existing typical rain and snow removal methods in single image: guidance image based image decomposition analysis, adaptive nonlocal means filter, and frequency-based analysis are also introduced in the literature. In this follows, we design a simple but effective method divide the rain or snow removal scheme into two parts, the first part is detection of rain or snow and the second part is inpainting. Besides, three existing typical underwater enhanced methods: histogram-based equalization, wavelength-based compensation, and fusion based are also introduced in the literature. In this follows, we design a simple but effective underwater enhanced method, and its main idea is combining the color correction, contrast stretching, and histogram equalization. Unlike other existing methods, we’ll get a better result which takes less processing time and highly enhances visibility and superior color fidelity by our method. We believe that we’ll run real-time on hardware in optimized circumstances.
19
Lin, Yu-Sheng, and 林育聖. "Adaptive Fast Image Dehazing Algorithms Based on Dark Channel Prior." Thesis, 2014. http://ndltd.ncl.edu.tw/handle/21503323826924481941.
Full textAbstract:
碩士朝陽科技大學
資訊工程系
102
The quality of digital images is easily affected by imaging conditions. Haze is one of adversarial conditions to degrade the image quality, such as contrast, visibility and color saturation. Many approaches have been proposed to deal with the hazy condition. However, most of them suffer from high computational complexity. In this thesis, several adaptive fast image dehazing algorithms based on the dark channel prior are presented. In the proposed dehazing algorithms, dark channel map is found through 1×1 minimum filter and then used to estimate the atmospheric light and transmission map. By this doing, the computation complexity is reduced and over-exposure problem generally happened in the dark channel based algorithms is avoided. Simulation results of several examples indicate that the proposed dehazing algorithms are generally able to obtain satisfactory dehazed images and are more efficient than the compared dehazing algorithms.
20
Chen, Mei-Ting, and 陳美庭. "Single Image Dehazing Based on Local and Global Airlight Estimation." Thesis, 2010. http://ndltd.ncl.edu.tw/handle/77728693037924241661.
Full text
21
Lai, Yi-Shuan, and 賴乙瑄. "Single Image Dehazing Using Transmission Heuristic with Optimal Transmission Map." Thesis, 2012. http://ndltd.ncl.edu.tw/handle/63146530493399830597.
Full textAbstract:
碩士國立清華大學
資訊工程學系
100
The poor visibility in bad weather condition, such as haze and fog, is caused by the stationary atmospheric effects of suspended particles. The challenge of restoring such atmospheric effects, usually referred to as “dehazing”, from single image mainly comes from the double uncertainty of scene depth and scene radiance. Approximation of the transmission, which encodes the scene depth information, is the most significant step to solve the dehazing problem. In this thesis, we propose to derive an optimal transmission map under a heuristic assumption in the dehazing model. The proposed objective function guarantees to have a global optimal solution, and the obtained transmission map is accurate and preserves the depth-consistency of the same object. Finally, we further take the difference in light wavelengths transmission between three color channels into account. Using the optimal transmission map and considering the different wavelengths of each color channel, our method recovers haze-free images with excellent result.
22
Chou, Chuan-Ju, and 周傳儒. "Single Image Dehazing Using Optimal Transmission Under Textured-Region Constraint." Thesis, 2013. http://ndltd.ncl.edu.tw/handle/34420618056321306945.
Full textAbstract:
碩士國立清華大學
資訊工程學系
101
Dehazing is an image restoration process to eliminate the hazy effect. Approximation of the transmission, which encodes the scene depth information, is the most significant step to tackle the dehazing problem. In this thesis, we propose a textured-region constraint to deal with inaccurate estimation on transmission within a textured region. The textured-region constraint assumes that transmission for pixels in a textured region should be similar. By including the textured-region constraint, the objective function guarantees to have a global optimal solution with constant textured-region transmission. Our experimental results show significant improvement in textured-region transmission and dehazed results.
23
KAUR, HARPREET. "STEP UP IN IMAGE DEHAZING USING WAVELET DECOMPOSITION & GUIDED FILTERS." Thesis, 2016. http://dspace.dtu.ac.in:8080/jspui/handle/repository/15390.
Full textAbstract:
In the present work, haze removal of images based on dark channel prior approach
has been proposed. Based on the observation that haze impacts low frequency
component of the image, only approximate part of the hazy image extracted using
lifting DB4 wavelet is processed by proposed dark channel prior algorithm. Haze
imaging model along with dark channel prior has been used to estimate atmospheric
light and transmission. Patch size used for determining dark channel has been made
adaptive on the basis of image size. Besides, airlight is obtained from larger patch to
ensure more accurate estimation in the presence of localized light sources. Further,
transmission has been estimated separately for each of the three color channels
considering the phenomena that different color channels undergo different scattering
in the atmosphere depending upon their wavelengths. In addition to this, transmission
of near white objects has been selectively increased to prevent them from getting over
saturated. Further, fast guided filter has been employed to refine transmission map in
place of soft matting. Results show that the proposed method gives better
performance as image contrast, entropy and SSIM have increased and runtime cost
has been significantly decreased. Problem of bluishness observed in classical method
also has been reduced.
24
LEE, HSIN-YU, and 李心瑜. "Adaptive Image Dehazing Technique based on Fusion Transmission and Sky Weight Detection." Thesis, 2019. http://ndltd.ncl.edu.tw/handle/f82ydf.
Full textAbstract:
碩士逢甲大學
資訊工程學系
107
Computer vision techniques are widely applied to the object detection, license plate recognition, remote sensing, and outdoor monitoring system. The performance of these applications mainly relies on the high quality of outdoor image. However, an outdoor image can be led to contrast decrease, color distortion, and unclear structure by poor weather conditions and human factors such as haze, fog, and air pollution. These issues may lower down the sharpness of a photo. Despite of the single-image dehazing is used to solve these issues, it cannot achieve a satisfactory result when the method deals with the bright scene and sky area. In this article, we aim to design an adaptive dehazing technique based on fusion transmission and sky weight detection. The sky weight detection is employed to distinguish the foreground and background, while detected results are applied to the fusion strategy to calculate deep and shallow transmissions. Thus, this can get rids of the subject of over-adjustment. Experimental results have demonstrated that the new method can outperform the latest state-of-the-art methods in terms of subjective and the objective assessments.
25
Lin, Cheng-Yang, and 林承洋. "Single Image Dehazing based on Modified Dark Channel Prior and Fog Density Detection." Thesis, 2012. http://ndltd.ncl.edu.tw/handle/87201645791502083261.
Full textAbstract:
碩士國立中山大學
電機工程學系研究所
100
In this thesis, a single image dehazing method based on modified dark channel prior and haze (fog) density detection is proposed. Dark channel prior dehazing algorithm is achieved good results for some haze images. However, we observed that haze images contain low and high haze density. Thus, the region of low haze density is unnecessary to dehaze. To solve this problem, we first defined the HSV distance, pixel-based dark channel prior and pixel-based bright channel prior to estimate the haze density. Further to enhance the dehazing performance of dark channel prior, the atmospheric light value and dehazing weighting is revised based on the HSV distance. Then the new transmission map is obtained. After that, a bilateral filter is applied to refine the transmission map, which can provide the higher accuracy of transmission map. Finally, the haze-free image is recovered by combining the input image and the refined transmission map. As a result, high-quality haze-free image can be recovered with lower computational complexity, which can be naturally extended to video dehazing.
26
Huang, I.-Chang, and 黃奕璋. "Hardware Architecture of Dehazing System for Enhancing Underwater Image with/without Extreme Contrast." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/h2ut4v.
Full textAbstract:
碩士國立中山大學
資訊工程學系研究所
106
The impurity in the water affects the clarity of underwater images, leading to great limitations on the application of underwater images in visual analysis and research. In addition to the fact that impurities cause image blurring, different refraction and scattering results are produced depending on the wavelength of the light, when light is transmitted in water,. The color contrast caused by the low contrast of the image and the partial color decay has similar results as the image taken in the fog. In recent years, an effective single image defogging algorithm, Dark Channel Prior, has been proposed in [1]. In many other studies (such as [2-7]), this algorithm is also applied to defogging underwater images. In addition the underexposed photos are often taken, when photographing in the underwater environment. Because the camera is like the human eye, it is not easy to clearly determine the outline, details, etc. of the object, when it receives the smooth, backlit and dark images at the same time. For the case of underexposed images, Pang et al. [8] proposed an enhanced algorithm for images with extremely low contrast. This method is based on the algorithm of Dong[9]. In addition to effectively enhancing the contrast of the image, it can also improve the halo caused by the algorithm of Dong[9]. Therefore, in order to improve these problems in underwater research analysis or underwater monitoring systems, a set of defogging systems that can perform image enhancement for underwater images with extreme contrast differences is indispensable. The defogging system proposed in this thesis contains four major steps: 1) With the RGB color model, the histogram distribution is used to determine whether it is an extreme contrast image or a normal contrast image. Accordingly the subsequent steps have different processing methods. 2) Since the underwater image and the image captured in the fog have similar problems such as contrast reduction and color shift, the dark channel prior can be used to calculate the atmospheric light source of the image. 3) In the dark channel prior, the calculation of the transfer rate is particularly important. Therefore, after estimating the transmission rate map of the image, we use a fast and simple guided filter to correct the edge of the transmission rate map to avoid the phenomenon of halation. 4) Finally, the image is restored by the method of defogging, and different absorption attenuation compensation is performed for the three color channels of RGB. To meet the real-time requirement, this thesis proposes a fast atmospheric light estimator hardware architecture based on the subsampling technique. As the result, the amount of data that needs to be read to estimate atmospheric light is greatly reduced, and the computational speed is much improved. Moreover, based on the algorithm of [8],this thesis also proposes the improved hardware architecture of the defogging system. In order to facilitate the implementation of the hardware architecture, the algorithm is simplified and the high performance and low cost can be achieved.
27
Wu, Bo-Yi, and 吳柏毅. "Design of a Haze Detection Algorithm and Hardware Implementation of an Image Dehazing Method." Thesis, 2015. http://ndltd.ncl.edu.tw/handle/5y558p.
Full textAbstract:
碩士國立雲林科技大學
電機工程系
103
This study aimed to explore how to remove haze from a single image effectively. On the hazy weather, we can't obtain the clearly and completely image of the original scene. Thus, how to effectively remove haze from images until the objects covered with haze can be recognized by human vision would be the issue for needed discussion in image processing. In the future, if the haze removal algorithm is practically applied to surveillance products as monitors, vehicle video recorders, cameras for other uses, and so on, the amounts of calculation will be taken into account besides the good image quality. Therefore, we designed a haze detection algorithm that could clearly identify if haze was in images. If the haze detection algorithm and the haze removal algorithm were simultaneously applied to the surveillance systems, the haze detection algorithm would be used to identify if haze was in images in advance. If so, the haze removal algorithm would be carried out, and then the images would be outputted after processing. If not, we could output the images directly without the haze removal algorithm. Hence, we could leave out the processing time of the haze removal algorithm for non-haze images to achieve the purpose of lowering the amounts of calculation. This study proposed the haze removal algorithm mainly divided into two steps: first, use the dark channel prior and the mean filter to measure atmospheric light; second, use dark channel prior and the patch center point to determine if an edge existed and obtain the transmission values. With the atmospheric light and transmission values, we could employ the formula to restore haze-free image. The haze detection algorithm was divided into two steps: first, use the Hue of HSV color model to analyze the images and then segment the sky in images; second, use dark channel prior to analyze histogram intensity for identifying whether there was haze in images. Based on the proposed haze removal algorithm, we also designed a comprehensive hardware structure. Furthermore, the Verilog development tool we used is Altera’s Quartus II software version 11.0 SP1. In view of hardware design, the dark channel prior techniques of the above two steps would be simultaneously scanned. That would be beneficial to decrease the time and increase the efficiency of employing the haze removal algorithm.
28
Shen, Kuan-Wei, and 沈冠緯. "Using a Hybrid of Interval Type-2 RFCMAC and Bilateral Filter for Satellite Image Dehazing." Thesis, 2017. http://ndltd.ncl.edu.tw/handle/764ug2.
Full textAbstract:
碩士國立勤益科技大學
資訊工程系
105
With advances in technology, the development of Remote Sensing Satellite Image has been real-time and accurate to monitor the environment of the surface or prevent the inevitable disaster earlier. Owing to the changeable weather is just like clouds or haze constituted by atmospheric particles, then this phenomenon cause the low contrast presented in satellite image and lose many information on the surface of the earth. Therefore, in this paper we propose an issue for dehazing to single satellite image, which is to enhance the contrast of image and filter the haze that cover the location, then the losing information will be back. First, we use Interval Type-2 RFCMAC Model to estimate the initial transmission map of the image. When facing the problems of halo and color over saturation, we adopt the bilateral filter and the quadratic function nonlinear transformation step by step to refine the initial transmission map. At the atmospheric light estimation, we adopt the first 1% brightest area as the color vector of atmospheric light. Finally, we take the refined transmission map and atmospheric light as the two parameters of reconstruct image. The experiment result shows that the method of satellite image dehazing has an effective results in visibility details and color contrast of reconstruction image. Furthermore, in order to prove the effective results, we take the visual assessment and quantitative evaluation respectively to compare with other authors. After visual assessment and quantitative evaluation, we get the better result in visual and data indeed.
29
Su, Yen-Chia, and 蘇衍嘉. "An end to end Single Image dehazing system based on Dense Block and Hybrid Loss Function." Thesis, 2019. http://ndltd.ncl.edu.tw/handle/m7bfe4.
Full textAbstract:
碩士國立臺灣科技大學
電機工程系
107
Image acquisition in a bad weather often results in visible distortions and color imbalance due to haze and imbalance light sources. In this scenario, the conventional dehazing algorithms try to estimate the transmission map or image prior. However, the condition of the transmission map is altered during different weather conditions and time. Consequently, the removal of haze with conventional strategies exist many limitations, and it needs improvements. In this thesis, a dehazing algorithm using the deep learning approach is proposed and the architecture is based on the U-Net. The model comprises of dense block networks, and incorporates different loss functions such as L1, L2, perceptual and total-variation. Dense Block is able to reuse feature map, and thus parameters can be reduced and also the network is more stable during training time. Compared with former dehazing algorithms, the input depth map is required to be estimated to acquire transmission rate map. Conversely, the proposed method only needs image pairs of the haze and haze-free images by applying supervised learning. This study also developed an effective deep learning architecture, combined with the concept of Deep DenseNet, which can extract important features of smog images and reconstruct these important features. In addition, we use a hybrid loss function combining their advantages, and then pick the appropriate weights to enhance the texture and detail of the reconstructed image. Finally, experimental results show that the deep learning based solution is significantly superior to the previous methods, and thus it can be a very practical method to address dehazing.
30
Lin, Chia-Hsiang, and 林嘉祥. "Nighttime Image Dehazing Based on Improved Erosion Dark Channel and Multi-scale Clipping Limit Histogram Equalization." Thesis, 2018. http://ndltd.ncl.edu.tw/handle/4xee4r.
Full textAbstract:
碩士國立臺灣科技大學
電機工程系
106
Photographs taken at night time with haze often suffers from poor visibility and color distortion due to uneven local light sources. In this scenario, the common daytime dehazing algorithm results increased atmospheric light due to the influence of local light sources. Due to this, the result of the transmission is affected, resulting in excessive haze removal, noise amplification and artifacts introduction. Consequently, this thesis proposes a new night time dehazing algorithm by modifying the calculation method of atmospheric light and transmission used in the dark channel priors proposed by He[2]. First, the proposed erosion Gaussian-based dark channel is applied to suppress local light sources. Subsequently, a modified transmission calculation method is proposed which combines the erosion operator with the transmission. As the transmission is obtained, the refinement operation is performed. The multi-scale guided filter utilizing improved transmission is applied to the proposed multi-scale guided filters for refinement. Through this improved calculations, the unpleasant issues can be well controlled, including excessive dehazing, excessive darkness, artifacts and excessive amplification of noise. In addition to the above three improvements on the dehazing algorithm, the ICLAHE proposed by Guo[22] et al is also modified with multi-scale clipping limits to further improve the image quality. With the proposed Multi-scale clipping limit of ICLAHE, details can be obtained. As documented in the experimental results, the proposed method can yield superior performance towards nighttime dehazing effect and of less computational complexity in comparison against the state-of-the-art methods.
31
PAZZAGLIA, FABIO. "Computer vision applied to underwater robotics." Doctoral thesis, 2016. http://hdl.handle.net/2158/1043212.
Full textAbstract:
Ocean and seafloors are today probably the less known and unexplored places on earth.Nowadays, the continuous technological improvements on underwater inspection offer new challenges and possibilities. Beside the lassic acoustic sensors, modern cameras are playing an ever increasing role in autonomous underwater navigation. In particular, The capability to perform a context-driven navigation, based on what the vehicle is actually seeing on the seafloor, is of great interest in many research fields, spanning from marine archaeology and biology to environment preservation. Industrial companies on oil and gas or submarine cabling, also have a strong interest in underwater robotics. The peculiarities of the underwater environment offer new opportunities to computer vision and pattern analysis researchers. This thesis analyses, discusses and extends computer vision techniques applied to the underwater environment. The main topic is the semantic classification of the seabed. A framework that may actually be embedded in an underwater vehicle and made to work in real time during the navigation was developed. The first part of this work addresses the problem of semantic image labelling. For this purpose a deep analysis of feature sets and related classification algorithms was carried out. The physical properties of light propagation in water need to be properly considered. Inspired by techniques for terrestrial single image dehazing, a new approach for underwater scenarios was developed. This approach is capable to significantly remove both the marine snow and the haze effects in images, and to effectively handle non-uniform and artificial lighting conditions. By jointly combining the results of underwater classification and the physical modelling of light transmission in water, a new feature set, more robust and with better discriminative performance was defined. Experimental results confirmed the accuracy improvements, Over the state-of-the-art obtained with the new feature set, in most critical environmental conditions. This work is largely based on original images and data, acquired during the European project ARROWS. The novelties introduced by this thesis may represent a basis for future applications, stimulating novel directions for research in computer vision and its applications to the underwater environment.