Dissertations / Theses on the topic 'Forest fires detection'

Every year wildfires threaten or destroy ecological habitats, man-made infrastructure and people’s lives. Additionally millions of dollars are spent each year trying to prevent and control these fires. Ideally if a wildfire can be detected before it rages out of control it can be extinguished and avoid large scale devastation. Traditional manned fire lookout towers are neither cost effective nor particularly efficient at detecting wildfire. It is proposed that temporal filtering can be used to isolate the signals created at the beginnings of potential wildfires. Temporal filtering can remove any background image and any periodic signals created by the camera movement. Once typical signals are analyzed, digital filters can be designed to pass fire signals while blocking the unwanted signals. The temporal filter passes only fire signals and signals generated by moving objects. These objects can be distinguished from each other by analyzing the objects mid and long wave energy profile. This algorithm is tested on 17 data sources and its results analyzed.

The aim of this study is to incorporate the spectral, temporal and spatial attributes of a smoke plume for Early Forest Fire Detection. Image processing techniques are used on multispectral (red, green, blue, mid-wave infrared, and long-wave infrared) video to segment and indentify the presence of a smoke plume within a scene. The temporal and spectral variance of a smoke plume is captured through Principal Component Analysis (PCA) where the Multispectral-Multitemporal PCA is performed on a sequence of video frames simultaneously. The presence of a plume existing in one of the higher order principal components is determined by the texture of its spatial content. The texture is characterized by statistical descriptors derived from the principal component‟s joint probability density distribution of intensities occurring within a spatial relationship, known as a Gray Level Co-Occurrence Matrix (GLCM). Initial analysis is performed on selected frames where only a subset of time is considered. Once the parameters are chosen from the static analysis, the algorithms are executed on video through time to validate the method. The results show that a smoke plume is readily segmented via PCA. Based on the five spectral bands over 3 seconds sampled at 1 second, the plume exists in the 7th principal component. Within these principal components, the smoke‟s presence is best identified by the correlation texture descriptor. The smoke is very spatially correlated compared to the scene at large. Therefore a spike in the spatial correlation of the principal components is all that is needed to identify the start of the smoke plume.

Forest fires constantly threaten ecological systems, infrastructure and human lives. The purpose behind this study is minimizing the devastating damage caused by forest fires. Since it is impossible to completely avoid their occurrences, it is essential to accomplish a fast and appropriate intervention to minimize their destructive consequences. The most traditional method for detecting forest fires is human based surveillance through lookout towers. However, this study presents a more modern technique. It utilizes land-based real-time multispectral video processing to identify and determine the possibility of fire occurring within the camera’s field of view. The temporal, spectral, and spatial signatures of the fire are exploited. The methods discussed include: (1) Range filtering followed by entropy filtering of the infrared (IR) video data, and (2) Principal Component Analysis of visible spectrum video data followed by motion analysis and adaptive intensity threshold. The two schemes presented are tailored to detect the fire core, and the smoke plume, respectively. Cooled Midwave Infrared (IR) camera is used to capture the heat distribution within the field of view. The fire core is then isolated using texture analysis techniques: first, range filtering applied on two consecutive IR frames, and then followed by entropy filtering of their absolute difference. Since smoke represents the earliest sign of fire, this study also explores multiple techniques for detecting smoke plumes in a given scene. The spatial and temporal variance of smoke plume is captured using temporal Principal Component Analysis, PCA. The results show that a smoke plume is readily segmented via PCA applied on the visible Blue band over 2 seconds sampled every 0.2 seconds. The smoke plume exists in the 2nd principal component, and is finally identified, segmented, and isolated, using either motion analysis or adaptive intensity threshold. Experimental results, obtained in this study, show that the proposed system can detect smoke effectively at a distance of approximately 832 meters with a low false-alarm rate and short reaction time. Applied, such system would achieve early forest fire detection minimizing fire damage. Keywords: Image Processing, Principal Component Analysis, PCA, Principal Component, PC, Texture Analysis, Motion Analysis, Multispectral, Visible, Cooled Midwave Infrared, Smoke Signature, Gaussian Mixture Model.

The goal of this study is to develop a smoke detecting algorithm using digital image processing techniques on multi-spectral (visible & infrared) video. By utilizing principal component analysis (PCA) followed by spatial filtering of principal component images the location of smoke can be accurately identified over a period of exposure time with a given frame capture rate. This result can be further analyzed with consideration of wind factor and fire detection range to determine if a fire is present within a scene. Infrared spectral data is shown to contribute little information concerning the smoke signature. Moreover, finalized processing techniques are focused on the blue spectral band as it is furthest away from the infrared spectral bands and because it experimentally yields the largest footprint in the processed principal component images in comparison to other spectral bands. A frame rate of .5 images/sec (1 image every 2 seconds) is determined to be the maximum such that temporal variance of smoke can be captured. The study also shows eigenvectors corresponding to the principal components that best represent smoke and are valuable indications of smoke temporal signature. Raw video data is taken through rigorous pre-processing schemes to align frames from respective spectral band both spatially and temporally. A multi-paradigm numerical computing program, MATLAB, is used to match the field of view across five spectral bands: Red, Green, Blue, Long-Wave Infrared, and Mid-Wave Infrared. Extracted frames are aligned temporally from key frames throughout the data capture. This alignment allows for more accurate digital processing for smoke signature. v Clustering analysis on RGB and HSV value systems reveal that color alone is not helpful to segment smoke. The feature values of trees and other false positives are shown to be too closely related to features of smoke for in solely one instance in time. A temporal principal component transform on the blue spectral band eliminates static false positives and emphasizes the temporal variance of moving smoke in images with higher order. A threshold adjustment is applied to a blurred blue principal component of non-unity principal component order and smoke results can be finalized using median filtering. These same processing techniques are applied to difference images as a more simple and traditional technique for identifying temporal variance and results are compared.

This thesis will analyze video from land-based, cooled mid-wave infrared cameras to identify temporal features indicative of a heat plume from a forest fire. Desirable features and methods will show an ability to distinguish between heat plume movement and other movements, such as foliage, vehicles, humans, and birds in flight. Features will be constructed primarily using combinations of statistics and principal component analysis (PCA) with intent to detect key characteristics of fire and heat plume: persistence and growth. Several classification systems will combine and filter the features in an attempt to classify pixels as either heat or non-heat. The classification systems will be tuned and compared with common metrics of error rate and computation time. It was found that the movement pattern of a heat plume could be distinguished from the similar movement pattern of foliage by detecting outlier movement patterns, a phenomenon associated with the growth property of fire. Outlier movement patterns were best detected by thresholding the quotient of mean and median of a set of variance measurements over time. The best tested classifier in terms of minimizing false positives without losing the heat signal came from PCA of a dual-range moving average difference.

Early detection and mitigation of wildfires can reduce devastating property damage, firefighting costs, pollution, and loss of life. This thesis proposes the method of Principal Component Analysis (PCA) of images in the temporal domain to identify a smoke plume in wildfires. Temporal PCA is an effective motion detector, and spatial filtering of the output Principal Component images can segment the smoke plume region. The effective use of other image processing techniques to identify smoke plumes and heat plumes are compared. The best attributes of smoke plume detectors and heat plume detectors are evaluated for combination in an improved wildfire detection system. PCA of visible blue images at an image sampling rate of 2 seconds per image effectively exploits a smoke plume signal. PCA of infrared images is the fundamental technique for exploiting a heat plume signal. A system architecture is proposed for the implementation of image processing techniques. The real-world deployment and usability are described for this system.

It is difficult to capture the early signs of a forest fire at night using current visible-spectrum sensor technology. Infrared (IR) light sensors, on the other hand, can detect heat plumes expelled at the initial stages of a forest fire around the clock. Long-wave IR (LWIR) is commonly referred to as the “thermal infrared” region where thermal emissions are captured without the need of, or reflections from, external radiation sources. Mid‑wave IR (MWIR) bands lie between the “thermal infrared” and “reflected infrared” (i.e. short-wave IR) regions. Both LWIR and MWIR spectral regions are able to detect thermal radiation; however, they differ significantly in regards to their detection sensitivity of forest-fire heat plumes. Fires fueled by organic material (i.e. wood, leaves, etc.) primarily emit hot carbon dioxide (CO2) gas at combustion. Consequently, because CO2 is also present in the atmosphere, re-emission restricts the spectral transmittance and hence spectral radiance over a wide range of frequencies in the MWIR region. Moreover, as the distance between the detector and fire’s heat plume becomes greater, the additional CO2 introduced into the detection path leads to further attenuation of photon emittance. Since these absorption frequencies also lie within the response bandwidth of the MWIR sensor material, captured heat plume radiation manifests itself as a group of “flooded” or saturated pixels that exhibit very little dynamic behavior. Meanwhile, since the LWIR spectral region is not significantly affected by atmospheric gas absorption, its sensor is able to capture the forest fire’s heat plume thermal signature at far range without such complications. By exploiting the underlying spectral differences between LWIR and MWIR regions, this study aims to achieve early forest fire heat plume detection via direct identification of its dynamic characteristics whist concurrent attenuation of detected non-fire-related radiation. A land‑based, co‑located, cooled-LWIR/cooled-MWIR (CLWIR/CMWIR) detector camera is used to capture and normalize synchronized video from which sequential spatial-domain difference frames are generated. Processed frames allow for effective extraction of the heat plume’s “flickering” features, which are characteristic to the early stages of a forest fire. A multilayer perceptron (MLP) neural network classifier is trained with feature points generated from known target samples (i.e. supervised learning). Resulting detection performance is assessed via detection time, error metrics, computation time, and parameter variation. Results indicate that heat plumes expelled at the early stages of a forest fire can be identified with high sensitivity, low false alarm, and at a farther range than commercial detectors.

A débit de données élevé, typiquement supérieur à 10 Gsymboles/s, les lignes de

télécommunication optique à fibre monomode souffrent de façon accrue des distorsions

inhérentes à la fibre et à l’architecture de transmission. Nous pouvons classer les

effets de fibre en plusieurs catégories:

– Les effets linéaires. La dispersion chromatique est entraînée par la dépendance en

fréquence de l’indice de réfraction de la fibre. Il en résulte un élargissement des

bits optiques. La dispersion des modes de polarisation prend son origine dans

la biréfringence de la fibre. La modélisation de cet effet est compliquée par son

caractère stochastique et variable dans le temps.

– Les effets non linéaires prennent leur origine dans un indice de réfraction de

fibre qui dépend du champ optique. Ces effets peuvent être classés en deux

catégories. Premièrement, les effets intérieurs à un canal dont le plus influant

est l’automodulation de phase qui découle de l’effet Kerr optique :l’intensité

d’une impulsion lumineuse influence sa propre propagation. Deuxièmement, il

existe des conséquences de l’effet Kerr par lesquelles les différents canaux, se

propageant au sein de la même fibre, s’influencent mutuellement. Le phénomène

le plus influent parmi ces derniers est la modulation de phase croisée :l’intensité

d’un canal influence la propagation dans un canal voisin.

– Les pertes par diffusion Rayleigh sont compensées par les amplificateurs distribués

le long de la ligne de transmission. L’amplification optique par l’intermédiaire

d’émission stimulée dans des dispositifs dopés aux ions Erbium est

accompagnée d’émission spontanée amplifiée. Ceci entraîne la présence d’un

bruit blanc gaussien se superposant au signal à transmettre.

– La gestion des canaux dans le réseau optique implique la présence dans les noeuds

du réseau de filtres de sélection, des multiplexeurs et démultiplexeurs.

Nous examinerons aussi les effets de ligne non inhérents à la fibre mais à l’architecture

de transmission. Les modèles de l’émetteur et du récepteur représentent les imperfections

d’implémentation des composants optiques et électroniques.

Un premier objectif est de définir et évaluer un format de modulation robuste aux

imperfections introduites sur le signal par la fibre optique et par l’émetteur/récepteur.

Deux caractéristiques fondamentales du format de modulation, determinants pour la

performance du système, sont étudiés dans ce travail :

– La forme d’ onde. Les symboles complexes d’information sont mis en forme par

un filtre passe-bas dont le profil influence la robustesse du signal vis-à-vis des

effets de ligne.

– La distribution des fréquences porteuses. Les canaux de communication sont

disposés sur une grille fréquentielle qui peut être définie de manière électronique

par traitement de signal, de manière optique ou dans une configuration hybride.

Lorsque des porteuses optiques sont utilisées, le bruit de phase relatif entre lasers

entraîne des effets d’ influence croisée entre canaux. En revanche, les limites des

implémentations électroniques sont données par la puissance des architectures

numériques.

Le deuxième objectif est de concevoir des techniques de traitement numérique du

signal implémentées après échantillonnage au récepteur afin de retrouver l’information

transmise. Les fonctions suivantes seront implémentées au récepteur :

– Les techniques d’estimation et d’égalisation des effets linéaires introduits par la

fibre optique et par l’émetteur et le récepteur. Le principe de l’égalisation dans

le domaine fréquentiel est de transformer le canal convolutif dans le domaine

temporel en un canal multiplicatif qui peut dès lors être compensé à une faible

complexité de calcul par des multiplications scalaires. Les blocs de symboles

émis doivent être rendus cycliques par l’ajout de redondance sous la forme d’un

préfixe cyclique ou d’une séquence d’apprentissage. Les techniques d’égalisation

seront comparées en termes de performance (taux d’erreurs binaires, efficacité

spectrale) et en termes de complexité de calcul. Ce dernier aspect est particulièrement

crucial en vue de l’optimisation de la consommation énergétique du

système conçu.

– Les techniques de synchronisation des signaux en temps/fréquence. Avant de

pouvoir égaliser les effets linéaires introduits dans la fibre, le signal reçu devra

être synchronisé en temps et en fréquence sur le signal envoyé. La synchronisation

est généralement accomplie en deux étapes principales :l’acquisition réalisée

avant de recevoir les symboles d’information don’t l’objectif est une première

estimation/compensation des effets de manière "grossière", le tracking réalisé en

parallèle à l’estimation des symboles d’information dont l’objectif est l’estimation

/compensation des effets de manière "fine". Les algorithmes d’acquisition et

de tracking peuvent nécessiter l’envoi d’informations connues du récepteur.

– Les techniques d’estimation et de compensation des imperfections de fonctionnement

de l’émetteur et du récepteur. Une structure de compensation des effets

introduits par les composants optiques et électroniques sera développée afin de

relâcher les contraintes d’implémentation de l’émetteur et du récepteur.

Etant donné la très haute cadence à laquelle les échantillons du signal sont produits

(plusieurs dizaines de Gech/s), une attention particulière est portée à la complexité de

calcul des algorithmes proposés.
Doctorat en Sciences de l'ingénieur
info:eu-repo/semantics/nonPublished

Les formats de modulation bidimensionnels (i.e. basés sur l’amplitude et la phase de l’onde porteuse) ont gagné depuis peu le domaine des transmissions par fibre optique grâce aux progrès conjoints de l’électronique rapide et du traitement du signal, indispensables pour réaliser les récepteurs opto-électroniques utilisant la détection cohérente des signaux optiques. Pour pallier les limites actuelles en rapidité de commutation des circuits intégrés électroniques, une voie de recherche a été ouverte il y a quelques années, consistant à utiliser des technologies optiques pour faciliter la parallélisation du traitement du signal, notamment dans l’étape d’échantillonnage ultra-rapide du signal rendu possible par des horloges optiques très performantes. Le thème principal de cette thèse concerne l’étude théorique et expérimentale de la fonction de conversion analogique-numérique (ADC) de signaux optiques par un récepteur opto-électronique cohérent, associant les étapes d’échantillonnage optique linéaire, de conversion analogique-numérique et de traitement du signal. Un prototype, utilisant une solution originale pour la source d’échantillonnage, est modélisé, réalisé et caractérisé, permettant la reconstruction temporelle de signaux optiques modulés selon divers formats : NRZ, QPSK, 16-QAM. Les limitations optiques et électroniques du système sont analysées, notamment l’impact sur la reconstruction des signaux de divers paramètres : le taux d’extinction de la source optique, les paramètres de l’ADC (bande passante BW, temps d’intégration et nombre effectif de bits ENOB). Par ailleurs, de nouveaux algorithmes de traitement du signal sont proposés dans le cadre de la transmission optique cohérente à haut débit utilisant des formats de modulation bidimensionnels (amplitude et phase) : deux solutions sont proposées pour la compensation du déséquilibre de quadrature IQ dans les transmissions mono-porteuses: une méthode originale de l’estimation du maximum du rapport signal sur bruit ainsi qu’une nouvelle structure de compensation et d’égalisation conjointes; ces deux méthodes sont validées expérimentalement et numériquement avec un signal 16-QAM. Par ailleurs, une solution améliorée de récupération de porteuse (décalage de fréquence et estimation de la phase), basée sur une décomposition harmonique circulaire de la fonction de maximum de vraisemblance logarithmique, est validée numériquement pour la première fois dans le contexte des transmissions optiques (jusqu’à une modulation de 128-QAM). Enfin les outils développés dans ce travail ont finalement permis la démonstration d’une transmission sur 100 km d’un signal QPSK à 10 Gbaud fortement limité par un bruit de phase non linéaire et régénéré optiquement à l’aide d’un limiteur de puissance préservant la phase basé sur une nanocavité de cristal photonique
Bi-dimensional modulation formats based on amplitude and phase signal modulation, are now commonly used in optical communications thanks to breakthroughs in the field of electronic and digital signal processing (DSP) required in coherent optical receivers. Photonic solutions could compensate for nowadays limitations of electrical circuits bandwidth by facilitating the signal processing parallelization. Photonic is particularly interesting for signal sampling thanks to available stable optical clocks. The heart of the present work concerns analog-to-digital conversion (ADC) as a key element in coherent detection. A prototype of linear optical sampling using an original solution for the optical sampling source, is built and validated with the successful equivalent time reconstruction of NRZ, QPSK and 16-QAM signals. Some optical and electrical limitations of the system are experimentally and numerically analyzed, notably the extinction ratio of the optical source or the ADC parameters (bandwidth, integration time, effective number of bits ENOB). Moreover, some new DSPs tools are developed for optical transmission using bi-dimensional modulation formats (amplitude and phase). Two solutions are proposed for IQ quadrature imbalance compensation in single carrier optical coherent transmission: an original method of maximum signal-to-noise ratio estimation (MSEM) and a new structure for joint compensation and equalization; these methods are experimentally and numerically validated with 16-QAM signals. Moreover, an improved solution for carrier recovery (frequency offset and phase estimation) based on a circular harmonic expansion of a maximum loglikelihood function is studied for the first time in the context of optical telecommunications. This solution which can operate with any kind of bi-dimensional modulation format signal is numerically validated up to 128-QAM. All the DSP tools developed in this work are finally used in a demonstration of a 10 Gbaud QPSK 100 km transmission experiment, featuring a strong non-linear phase noise limitation and regenerated using a phase preserving and power limiting function based on a photonic crystal nanocavity

Forest fires are an increasingly relevant problem nowadays with the worsening of global warming’s most severe consequences. These fire occurrences, that can cause immense damage to forest ecosystems and have a great negative impact in peoples lives,begin often with rekindles. These problems can be very difficult to tackle, needing to involve a lot of people to surveil the areas at risk. A system capable of executing this surveillance protocol and alerting the fire fighting authorities of fire and possible rekindle occurrences would be extremely beneficial in these scenarios.A system aiming to achieve this goal is being implemented, composed of an UAV equipped with a multispectral camera, capturing aerial images of these areas. This dissertation presents a fire detection model to be used in prescribed fires and rekindling situations, identifying fire instances within the captured images. It makes use of the camera’s various spectral bands to highlight the areas at greatest risk and of deep learning technology to autonomously recognise these areas.
Incêndios florestais são um problema cada vez mais relevante nos dias de hoje com o agravamento das consequências mais graves do aquecimento global. Estas ocorrências,que podem causar imensos danos aos ecossistemas florestais e ter um grande impacto negativo na vida das pessoas, são muitas vezes iniciadas por reacendimentos. Estes problemas podem ser muito difíceis de combater, necessitando de envolver muitas pessoas para vigiar as áreas de risco. Um sistema capaz de executar este protocolo de vigilância e alertar as autoridades de combate a incêndio sobre fogos e possíveis reacendimentos seria extremamente benéfico nestes cenários.Para alcançar este objetivo, está a ser implementado um sistema composto por um UAV, equipado com uma câmera multiespectral, que irá capturar imagens aéreas dessas áreas. Esta dissertação apresenta um modelo de detecção de incêndios para ser utilizado em situações de fogos controlados e reacendimentos, identificando ocorrências de fogo nas imagens capturadas. Faz uso das várias bandas espectrais da câmera para destacar as áreas de maior risco e de tecnologia de aprendizagem automática para reconhecer essas áreas de forma autônoma.

Dissertation presented as the partial requirement for obtaining a Master's degree in Geographic Information Systems and Science
Nesta dissertação apresenta-se um estudo sobre a influência da resolução espacial na detecção de alterações do coberto florestal. Numa revisão bibliográfica analisa-se os princípios da radiação eletromagnética, o comportamento espectral das principais alterações do coberto florestal (áreas ardidas, cortes florestais e novas plantações), os métodos de detecção de alterações baseados em dados de observação da Terra e os efeitos da escala na extracção deste tipo de informação. Após a selecção de uma metodologia de detecção de alterações, procura-se identificar as principais alterações do coberto florestal ocorridas em Portugal Continental entre os Outonos de 2002 e 2003. Esta metodologia é aplicada a três imagens com diferentes resoluções espaciais (250, 500 e 1000 metros) e a todas as combinações possíveis de um conjunto de parâmetros (limiar de detecção, unidade mínima cartográfica e máscaras de análise). Os resultados são analisados com base em índices de avaliação de qualidade (erros de comissão e de omissão), através de comparação com dados de referência. Analisa-se os efeitos das diferentes resoluções espaciais nos resultados desta detecção.(...)

碩士
國立暨南國際大學
資訊工程學系
101
Along with the progress of computer computation capabilities, sophisticated image processing/understanding methods have been developed and the functions of intelligent video surveillance systems have been greatly extended. In this thesis, we develop a video-based fire and smoke detection system based on the random forest algorithm. We use the distinct color and image variation properties of fire/smoke to select candidate regions. Then, image features of texture and motion patterns of the candidate regions are analyzed to determine any fire/smoke region. We propose to extract the features of both the texture and motion patterns of the fire/smoke with the local binary pattern (LBP) method. The random forest method is augmented to use the LBP features for fire/smoke detection to reduce false positive and enhance the fire and smoke detection rate.

The main purpose for this project is the development of a semi-autonomous wireless sensor network for fire detection in remote territory. Making use of the IEEE 802.15.4 standard, a wireless standard for low-power, low-rate wireless sensor networks, a real sensor network and web application will be developed and deployed with the ability to monitor sensor data, detect a fire occurrence and generate early fire alerts.

The main purpose for this project is the development of a semi-autonomous wireless sensor network for fire detection in remote territory. Making use of the IEEE 802.15.4 standard, a wireless standard for low-power, low-rate wireless sensor networks, a real sensor network and web application will be developed and deployed with the ability to monitor sensor data, detect a fire occurrence and generate early fire alerts.

碩士
國立屏東科技大學
森林系
92
The most important act to prevent the occurrence of forest fire is familiar with the factors cause to forest fire. It is necessary to identify the forest fire potential danger area in order to notice the manager to enhance the times of patrol, and under control whole condition when the forest fire occurring. This study aims on how to use environment data and satellite image to Detection forest fire in Taiwan forest. The first part of this study is using environmental data and Geographic Information System (GIS) techniques to construct the criteria of forest fire potential danger area model. Environmental data include elevation, slope, aspect, far away from the road, average relative humidity of years, average temperature of years, average rainfall of years and whole kind of forest vegetation types in Taiwan. The second parts of this study is using MODIS image to calculate Normalized Difference Vegetation Index (NDVI), Weighted Difference Vegetation Index (WDVI), Normalized Thermal Index (NTI) and Normalized Difference Water Index (NDWI), and with those indexes to construct forest fire Detection model. The results of this study are fellows: 1.Through with the analysis of forest fire data of 2003, the most dangerous period of forest fire occurring is between Novembers to March in the whole year. For the susceptive zone of forest fire, the range within 2,000 meters distant from road, under mean sea level 500 meter, slope between 26-35 degrees, aspect are southwest and south, and vegetation type are grasslands and low tree density forest. For climate, the average rainfall of years under 2,000 millimeters, the average temperature of years upward 21℃, the average relative humidity of years under 79 %, also the susceptive zone of forest fire. 2.This research constructed the criteria of forest fire potential danger area model with those environmental data. The more the sums of criteria score the higher probability to occur forest fire. There are divided 5 grades to identify the forest fire potential level. 3.Using WDVI, NDVI and NTI to construct forest fire Detection model, but NDWI was exclusive in the model because it is not significant difference after the statistical analysis. 4.Band 21 of MODIS image was designed for detect the temperature of land surface and forest fire. Examine with the experimental data it seems would not affected by the cloud in detect the temperature of land surface. And from the result shows that it is feasible to apply in Detection forest fire in Taiwan.

Dissertação de Mestrado Integrado em Engenharia Electrotécnica e de Computadores apresentada à Faculdade de Ciências e Tecnologia
Os fogos florestais são dos desastres naturais mais preocupantes que aconteceram pelo mundo inteiro nos últimos anos, tendo queimado centenas de milhares de quilómetros quadrados de áreas rurais e urbanas, causado centenas de vítimas mortais e milhares de milhões de euros em prejuízos.Esta dissertação introduz uma rede sensorial de baixo custo e inteligente, baseada em sensores térmicos matriciais com campo de visão estreito, que consegue detetar e localizar fontes de calor, como incêndios florestais, em tempo real. Este sistema consegue detetar incêndios florestais em aproximação e foi desenvolvido para ser usado em combinação com mecanismos ativos de proteção contra fogo. Um destes mecanismos é uma cerca para proteção contra incêndios florestais, que combina um sistema de aspersores com uma tela térmica, para tentar sustentar o avanço da frente de chama. Este sistema é também capaz de enviar alarmes, sempre que um incêndio florestal é detetado, para alertar entidades de proteção civil e proprietários de terrenos. Integrado neste sistema está também um sensor ambiental, capaz de medir temperatura ambiente, humidade relativa, pressão atmosférica e compostos orgânicos voláteis (VOC). Estes dados podem ser utilizados para geral informação sobre o risco de incêndio florestal, usando o Índice Meteorológico de Perigo de Incêndio Florestal (FWI). O principal objetivo deste trabalho é desenvolver uma solução que proactivamente proteja pequenas áreas de interesse: casas isoladas, pequenas aldeias, parques de campismo, festivais, etc. Testes usando pequenas frentes de fogo validaram a solução proposta para detetar e localizar a exitência de fogo a curta distância.
Forest fires are one of the most worrying natural disasters that happened around the world in the last few years having burned hundreds of thousands of square kilometres of rural and urban areas, having caused hundreds of deaths and billions of euros in losses.This dissertation introduces a novel low-cost smart system of networked nodes, based on narrow Field of View (FOV) Far Infared (FIR) Thermal Array sensors, which can detect and locate strong heat sources, such as forest fires, in real-time. This system is able to detect approaching wildfires, and was designed to be used in combination with active fire protection mechanisms. One of these mechanisms, is a fence for protection against forest fires, which combines a water sprinkler system with a thermal screen, in order to sustain or delay the advance of the flame front. This system is also capable of emitting alarms, when forest fires are detected, to alert civil protection entities and rural land owners. The system includes an environmental sensor, capable of sensing ambient temperature, relative humidity, barometric pressure, and Volatile Organic Compounds (VOCs). This data could serve a source of fire risk information, coupled with the Fire Weather Index (FWI). The prime objective of this work was to develop a solution to proactively protect small areas of interest such as: isolated homes, small villages, camping sites, music festivals, etc when a fire is within 50 meters. Tests conducted with real fire fronts validated the proposed solution for detecting and pinpointing fire existence at close range.

Dissertação de Mestrado Integrado em Engenharia Mecânica apresentada à Faculdade de Ciências e Tecnologia
Forest fires are a very present reality in the Portuguese panorama, possibly being thenatural disaster that most affects the country, both in environmental terms, as well as ineconomic and social terms. Firefighting is a very dangerous type of activity that sometimestakes the life of those who combat it. To reduce the risk to which firefighters are subject, itis necessary to develop new strategies and techniques that preserve the physical integrity offirefighter. For this, it makes sense to resort to technological advances already made to tryto develop a solution that helps firefighters fight fires without constantly being in danger.This work consists of the development of an autonomous targeting system that mustdetect and locate hotspots through an IR sensor, and that will be incorporated in a drone thatmust work as autonomously as possible. The main objectives of this work are thedevelopment of an algorithm that determines the location of the hotspot, the study of thewind effect on the waterjet that will hit the hotspot, and the development of an algorithmthat determines the correct movements so that the waterjet hits accurately on the hotspot.For the development of this work, the IR sensor Flir Lepton 3.5 was used, whoseobjective is to generate a matrix of values that indicate the temperature in each pixel. It willbe this sensor that will indicate the position in which the hotspot is located. Another keysensor is the PX4Flow. This sensor has a sonar attached that allows the user to knowaccurately how high the drone is from the ground. This sensor is very important in definingthe trajectory of the waterjet. There is also a flight controller, which is the Pixhawk 4 mini,and which is responsible for indicating parameters such as the “yaw” angle at which thedrone is. All information is sent to the Adafruit Huzzah32 microcontroller. This device isresponsible for communication via Wi-Fi between the drone and the user.The first steps of this work consisted of determining the distance at which the hotspotwas from the drone, according to the characteristics of the matrix generated by the IR sensor.Then the trajectory of the waterjet was determined. For this, it was necessary to analyze theeffect of airflow on a waterjet in a wind tunnel to determine its aerodynamic characteristics.Knowing the distance at which the hotspot is from the drone and knowing the trajectory ofthe waterjet, it is possible to know the movement that the drone must do to be aligned withthe hotspot. The next step of this work consisted of making the communication between thesensors and controllers. For this, communication protocols, like I2C and SPI, were used. AC# interface was also developed that allows the user to remotely control the drone and, at the same time, receive data via Wi-Fi in real-time for all the required parameters. The laststep of this work consisted of conducting experimental tests on this entire system installedin the drone, to verify if the entire theoretical approach was done correctly.
Os incêndios florestais são uma realidade bem presente no panorama português,sendo possivelmente a catástrofe natural que mais afeta o país, quer em termos ambientais,como económicos e sociais. O combate aos incêndios constitui um tipo de atividade muitoperigosa, e que por vezes tira a vida a quem a realiza. Na tentativa de reduzir o risco a queos bombeiros estão sujeitos, é necessário desenvolver novas estratégias e técnicas quepermitam preservar a integridade física dos bombeiros. Para isso faz sentido recorrer aosavanços tecnológicos já feitos para tentar desenvolver uma solução que auxilie os bombeirosno combate aos incêndios sem que estes estejam constantemente sob perigo.Este trabalho consiste no desenvolvimento de um sistema de mira automática quedeve detetar e localizar pontos quentes através de um sensor IR, e que será incorporado numdrone que deve funcionar da forma mais autónoma possível. Os principais objetivos destetrabalho são o desenvolvimento de um algoritmo que determine a localização do pontoquente, o estudo do efeito do vento no jato de água que irá atingir o ponto quente, e odesenvolvimento de um algoritmo que determine o ângulo “yaw” do drone e o ângulo “tilt”da agulheta de forma que o jato de água incida sobre o ponto quente.Para o desenvolvimento deste trabalho foi utilizado o sensor IR Flir Lepton 3.5, cujoobjetivo é gerar uma matriz de valores que indiquem a temperatura em cada pixel. Será estesensor que indicará qual a posição em que se encontra o ponto quente. Outro sensorfundamental é o PX4Flow. Este sensor tem acoplado a si um sonar que permite ao utilizadorsaber de forma muito aproximada a que altura se encontra o drone do solo. Este sensor émuito importante na definição da trajetória do jato de água. Existe ainda um controlador devoo, que é o Pixhawk 4 mini e que é responsável por indicar parâmetros tais como o ângulo“yaw” a que o drone se encontra. Toda a informação é enviada para o microcontroladorAdafruit Huzzah32. Este dispositivo é responsável pela comunicação via Wi-Fi entre o dronee o utilizador.As primeiras etapas deste trabalho consistiram em determinar a distância a que oponto quente se encontrava do drone, de acordo com as características da matriz gerada pelosensor IR. De seguida foi determinada a trajetória do jato de água. Para isso foi necessárioanalisar o efeito de um escoamento de ar sobre um jato de água num túnel de vento de forma a determinar as suas características aerodinâmicas. Sabendo a distância a que o ponto quenteestá do drone e sabendo a trajetória do jato de água, é possível saber qual o movimento queo drone tem de realizar para se enquadrar com o ponto quente. A etapa seguinte destetrabalho consistiu em fazer a comunicação entre os sensores e os controladores. Foi aindadesenvolvida uma interface em C# que permite ao utilizador controlar remotamente o dronee, ao mesmo tempo, receber dados via Wi-Fi em tempo real de todos os parâmetrosrequeridos. A última etapa deste trabalho consistiu em realizar ensaios experimentais a todoeste sistema instalado no drone, de forma a verificar se toda a abordagem teórica foi feita deforma correta.
Outro - European Fund for Regional Development through national funds (Compete2020 and Portugal2020)

In recent years landscape-scale fires have occurred in mainland Southeast Asia, including important protected areas (PAs). There has been increasing concern that landscape-scale fires are degrading the seasonally dry evergreen forest (SEF) element of the forest mosaic to more open deciduous forest and savanna, with serious implications for biodiversity conservation. Present management approaches, including fire suppression and prescribed burning, have not been effective managing for landscape-scale fire. Research was undertaken to investigate the occurrence, cause, effect, frequency and predictability of fire in SEF. SEF has the greatest species biodiversity in the forest mosaic and is potentially the most affected by fire, yet little research has been done on fire in SEF in mainland Southeast Asia. Huai Kha Khaeng (HKK) Wildlife Sanctuary in Thailand was selected as the study area. The objectives included: 1) investigate the area of SEF burned in HKK from 1988 to 2002; 2) investigate the conditions for fire in SEF; 3) determine whether the area of SEF in HKK declined as a result of fire; 4) determine the frequency of fire season years between 1984 and 2001 with the conditions for fire spread in SEF; and 5) determine whether there is a significant relationship between pre-fire season drought codes (Keetch-Byram Drought Index (KBDI) and Canadian Drought Code (DC)) and identified SEF fire season years for 1981 to 2003. Methods included: development of a Landsat fire history with associated interviews and reconnaissance field checks; fieldwork lighting test fires and measuring fuel characteristics; remote sensing change detection work using Landsat imagery; generation of a twenty-one year daily relative humidity minimum record for SEF; and logistic regression of the pre-fire season drought code values with identified SEF ‘fire’ and ‘non-fire’ years. Results showed: 1. Extensive areas of SEF have burned, but that Landsat imagery was not suitable for detecting fire in intact SEF. 2. SEF burned in years when there were fires burning adjacent to SEF in mid March and the moisture content of the SEF leaf litter fuel was less than 15%. 3. Fifteen percent of SEF in HKK has been either degraded or converted to deciduous forest forms in 12 years. 4. Conditions for fire spread in SEF occurred four times in 17 consecutive years. 5. A significant relationship exists between both the Keetch-Byram Drought Code (KBDI) and Canadian Drought Code (DC) and the SEF fire years. Implications are that large-scale fires have adversely affected intact SEF in HKK, and that the current damaging situation can be expected to continue. Whereas the extent of burning in intact SEF is not known, the need to manage the situation is immediate. Landscape-scale fires in HKK can be managed by using January 31st drought code values to predict potential large-scale fire years, followed by an aggressive fire suppression campaign in those years. In other years, fires can be allowed to burn without serious threat to the forest mosaic, and should to some extent be encouraged to maintain open deciduous forests and savanna. Additional research is required to determine whether a similar approach can be used for protected areas in other parts of the region.

The research problem that this thesis seeks to examine is a method of predicting conventional fire hazards using data drawn from specific regions, namely the Sooke and Goldstream watershed regions in coastal British Columbia. This thesis investigates whether LiDAR data can be used to describe conventional forest stand fire hazard classes. Three objectives guided this thesis: to discuss the variables associated with fire hazard, specifically the distribution and makeup of fuel; to examine the relationship between derived LiDAR biometrics and forest attributes related to hazard assessment factors defined by the Capitol Regional District (CRD); and to assess the viability of the LiDAR biometric decision tree in the CRD based on current frameworks for use. The research method uses quantitative datasets to assess the optimal generalization of these types of fire hazard data through discriminant analysis. Findings illustrate significant LiDAR-derived data limitations, and reflect the literature in that flawed field application of data modelling techniques has led to a disconnect between the ways in which fire hazard models have been intended to be used by scholars and the ways in which they are used by those tasked with prevention of forest fires. It can be concluded that a significant tradeoff exists between computational requirements for wildfire simulation models and the algorithms commonly used by field teams to apply these models with remote sensing data, and that CRD forest management practices would need to change to incorporate a decision tree model in order to decrease risk.
Graduate
0799
0478
christos@koulas.ca

Neste projecto é testada uma metodologia para actualização de mapas de ocupação do solo já existentes, derivados de fotografia aérea, usando uma imagem satélite para posterior modelação com vista à obtenção da cartografia do risco de incêndio actualizada. Os diferentes passos metodológicos na fase da actualização dos mapas de ocupação de solo são: Classificação digital das novas ocupações, Produção do mapa de alterações, Integração de informação auxiliar, Actualização da Cartografia Temática. Para a produção do mapa de alterações a detecção de alterações foi efectuada através de expressões de Álgebra de Mapas. A classificação digital foi realizada com um classificador assistido - Classificador da Máxima Verosimilhança. A integração de informação auxiliar serviu para melhorar os resultados da classificação digital, nomeadamente em termos das áreas ardidas permitindo uma resolução temática mais detalhada. A actualização resultou da sobreposição do mapa das áreas alteradas classificadas com o mapa desactualizado. Como produto obteve-se a Carta de Alterações da Ocupação do Solo com Potencial Influência no Risco de Incêndio actualizada para 2008, base para a fase da Modelação do Risco. A metodologia foi testada no concelho de Viseu, Centro de Portugal. A Carta de Uso e Ocupação do Solo de Portugal Continental para 2007 (COS2007) foi utilizada como carta de referência. A nova carta actualizada para 2008, no concelho de Viseu, apresenta 103 classes temáticas, 1ha de unidade mínima e 90% de precisão global. A modelação do risco de incêndio geralmente é feita através de índices que variam, de forma geral, numa escala qualitativa, tendo como fim possibilitar a definição de acções de planeamento e ordenamento florestal no âmbito da defesa da floresta contra incêndios. Desta forma, as cartas de risco são indicadas para acções de prevenção, devendo ser utilizadas em conjunto com a carta da perigosidade que juntas podem ser utilizadas em acções de planeamento, em acções de combate e supressão. A metodologia testada, neste projecto, para elaboração de cartografia de risco foi, a proposta por Verde (2008) e adoptada pela AFN (2012). Os resultados apresentados vão precisamente ao encontro do que diz respeito no Guia Técnico para Plano Municipal de Defesa da Floresta Contra Incêndios, "O mapa de Risco combina as componentes do mapa de perigosidade com as componentes do dano potencial (vulnerabilidade e valor) para indicar qual o potencial de perda face ao incêndio florestal".

During the summer, forest fires are the main reason for deforestation and the damage caused to homes and property in different communities around the world. The use of Unmanned Aerial Vehicles (UAVs, and also known as drones) applications has increased in recent years, making them an excellent solution for difficult tasks such as wildlife conservation and forest fire prevention. A forest fire detection system can be an answer to these tasks. Using a visual camera and a Convolutional Neural Network (CNN) for image processing with an UAV can result in an efficient fire detection system. However, in order to be able to have a fully autonomous system, without human intervention, for 24-hour fire observation and detection in a given geographical area, it requires a platform and automatic recharging procedures. This dissertation combines the use of technologies such as CNNs, Real Time Kinematics (RTK) and Wireless Power Transfer (WPT) with an on-board computer and software, resulting in a fully automated system to make forest surveillance more efficient and, in doing so, reallocating human resources to other locations where they are most needed.
Durante o verão, os incêndios florestais constituem a principal razão do desflorestamento e dos danos causados às casas e aos bens das diferentes comunidades de todo o mundo. A utilização de veículos aéreos não tripulados (VANTs), em inglês denominados por Unmanned Aerial Vehicles (UAVs) ou Drones, aumentou nos últimos anos, tornando-os uma excelente solução para tarefas difíceis como a conservação da vida selvagem e prevenção de incêndios florestais. Um sistema de deteção de incêndio florestal pode ser uma resposta para essas tarefas. Com a utilização de uma câmara visual e uma Rede Neuronal Convolucional (RNC) para processamento de imagem com um UAV pode resultar num eficiente sistema de deteção de incêndio. No entanto, para que seja possível ter um sistema completamente autónomo, sem intervenção humana, para observação e deteção de incêndios durante 24 horas, numa dada área geográfica, requer uma plataforma e procedimentos de recarga automática. Esta dissertação reúne o uso de tecnologias como RNCs, posicionamento cinemático em tempo real (RTK) e transferência de energia sem fios (WPT) com um computador e software de bordo, resultando num sistema totalmente automatizado para tornar a vigilância florestal mais eficiente e, ao fazê-lo, realocando recursos humanos para outros locais, onde estes são mais necessários.