The Niagara Escarpment is one of Southern Ontario’s most important landscapes. Due to the nature of the landform and its location, the Escarpment is subject to various development pressures including urban expansion, mineral resource extraction, agricultural practices and recreation. In 1985, Canada’s first large scale environmentally based land use plan was put in place to ensure that only development that is compatible with the Escarpment occurred within the Niagara Escarpment Plan (NEP). The southern extent of the NEP is of particular interest in this study, since a portion of the Plan is located within the rapidly expanding Greater Toronto Area (GTA). The Plan area located in the Regional Municipalities of Hamilton and Halton represent both urban and rural geographical areas respectively, and are both experiencing development pressures and subsequent changes in land cover. Monitoring initiatives on the NEP have been established, but have done little to identify consistent techniques for monitoring land cover on the Niagara Escarpment. Land cover information is an important part of planning and environmental monitoring initiatives. Remote sensing has the potential to provide frequent and accurate land cover information over various spatial scales. The goal of this research was to examine land cover change in the Regional Municipalities of Hamilton and Halton portions of the NEP. This was achieved through the creation of land cover maps for each region using Landsat 5 Thematic Mapper (TM) remotely sensed data. These maps aided in determining the qualitative and quantitative changes that had occurred in the Plan area over a 20 year time period from 1986 to 2006. Change was also examined based on the NEP’s land use designations, to determine if the Plan policy has been effective in protecting the Escarpment. To obtain land cover maps, five different supervised classification methods were explored: Minimum Distance, Mahalanobis Distance, Maximum Likelihood, Object-oriented and Support Vector Machine. Seven land cover classes were mapped (forest, water, recreation, bare agricultural fields, vegetated agricultural fields, urban and mineral resource extraction areas) at a regional scale. SVM proved most successful at mapping land cover on the Escarpment, providing classification maps with an average accuracy of 86.7%. Land cover change analysis showed promising results with an increase in the forested class and only slight increases to the urban and mineral resource extraction classes. Negatively, there was a decrease in agricultural land overall. An examination of land cover change based on the NEP land use designations showed little change, other than change that is regulated under Plan policies, proving the success of the NEP for protecting vital Escarpment lands insofar as this can be revealed through remote sensing. Land cover should be monitored in the NEP consistently over time to ensure changes in the Plan area are compatible with the Niagara Escarpment. Remote sensing is a tool that can provide this information to the Niagara Escarpment Commission (NEC) in a timely, comprehensive and cost-effective way. The information gained from remotely sensed data can aid in environmental monitoring and policy planning into the future.

Wetlands are considered to be one of the most valuable natural occurring forms of land cover in the world. Hydrologic regulation, carbon sequestration, and habitat provision for a wide assortment of flora and fauna are just a few of the benefits associated with wetlands. The implementation of satellite remote sensing has been demonstrated to be a reliable approach to monitoring wetlands over time. Unfortunately, a national wetland inventory does not exist for Canada at this time. This study employs a supervised classification method of Landsat satellite imagery between 1976 and 2008 within the Tunkwa watershed, southwest of Kamloops, British Columbia, Canada. Images from 2005 and 2008 were repaired using a gap-filling technique due to do the failure of the scan-line corrector on the Landsat 7 satellite in 2003. Percentage pixel counts for wetlands were compared, and a diminishing trend was identified; approximately 4.8% of wetland coverage loss was recognized. The influence of the expansion of Highland Valley Copper and the forestry industry in the area may be the leading causes of wetland desiccation. This study expresses the feasibility of wetland monitoring using remote sensing and emphasizes the need for future work to compile a Canadian wetland inventory.

This project analyzes satellite images from 1976, 1985 and 2000 of Dutse, Jigawa state, in north-western Nigeria. The analyzed satellite images were used to determine land-use and vegetation changes that have occurred in the land-use from 1976 to 2000 will help recommend possible planning measures in order to protect the vegetation from further deterioration. Studying land-use change in north-western Nigeria is essential for analyzing various ecological and developmental consequences over time. The north-western region of Nigeria is of great environmental and economic importance having land cover rich in agricultural production and livestock grazing. The increase of population over time has affected the land-use and hence agricultural and livestock production. On completion of this project, the possible land use changes that have taken place in Dutse will be analyzed for future recommendation. The use of supervised classification and change detection of satellite images have produced an economic way to quantify different types of landuse and changes that has occurred over time. The percentage difference in land-use between 1976 and 2000 was 37%, which is considered to be high land-use change within the period of study. The result in this project is being used to propose planning strategies that could help in planning sustainable land-use and diversity in Dutse.

Today's high-throughput data collection devices, e.g. spectrometers and gene chips, create information in abundance. However, this poses serious statistical challenges, as the number of features is usually much larger than the number of observed units.  Further, in this high-dimensional setting, only a small fraction of the features are likely to be informative for any specific project. In this thesis, three different approaches to the two-class supervised classification in this high-dimensional, low sample setting are considered. There are classifiers that are known to mitigate the issues of high-dimensionality, e.g. distance-based classifiers such as Naive Bayes. However, these classifiers are often computationally intensive and therefore less time-consuming for discrete data. Hence, continuous features are often transformed into discrete features. In the first paper, a discretization algorithm suitable for high-dimensional data is suggested and compared with other discretization approaches. Further, the effect of discretization on misclassification probability in high-dimensional setting is evaluated.   Linear classifiers are more stable which motivate adjusting the linear discriminant procedure to high-dimensional setting. In the second paper, a two-stage estimation procedure of the inverse covariance matrix, applying Lasso-based regularization and Cuthill-McKee ordering is suggested. The estimation gives a block-diagonal approximation of the covariance matrix which in turn leads to an additive classifier. In the third paper, an asymptotic framework that represents sparse and weak block models is derived and a technique for block-wise feature selection is proposed.      Probabilistic classifiers have the advantage of providing the probability of membership in each class for new observations rather than simply assigning to a class. In the fourth paper, a method is developed for constructing a Bayesian predictive classifier. Given the block-diagonal covariance matrix, the resulting Bayesian predictive and marginal classifier provides an efficient solution to the high-dimensional problem by splitting it into smaller tractable problems. The relevance and benefits of the proposed methods are illustrated using both simulated and real data. / Med dagens teknik, till exempel spektrometer och genchips, alstras data i stora mängder. Detta överflöd av data är inte bara till fördel utan orsakar även vissa problem, vanligtvis är antalet variabler (p) betydligt fler än antalet observation (n). Detta ger så kallat högdimensionella data vilket kräver nya statistiska metoder, då de traditionella metoderna är utvecklade för den omvända situationen (p<n).  Dessutom är det vanligtvis väldigt få av alla dessa variabler som är relevanta för något givet projekt och styrkan på informationen hos de relevanta variablerna är ofta svag. Därav brukar denna typ av data benämnas som gles och svag (sparse and weak). Vanligtvis brukar identifiering av de relevanta variablerna liknas vid att hitta en nål i en höstack. Denna avhandling tar upp tre olika sätt att klassificera i denna typ av högdimensionella data.  Där klassificera innebär, att genom ha tillgång till ett dataset med både förklaringsvariabler och en utfallsvariabel, lära en funktion eller algoritm hur den skall kunna förutspå utfallsvariabeln baserat på endast förklaringsvariablerna. Den typ av riktiga data som används i avhandlingen är microarrays, det är cellprov som visar aktivitet hos generna i cellen. Målet med klassificeringen är att med hjälp av variationen i aktivitet hos de tusentals gener (förklaringsvariablerna) avgöra huruvida cellprovet kommer från cancervävnad eller normalvävnad (utfallsvariabeln). Det finns klassificeringsmetoder som kan hantera högdimensionella data men dessa är ofta beräkningsintensiva, därav fungera de ofta bättre för diskreta data. Genom att transformera kontinuerliga variabler till diskreta (diskretisera) kan beräkningstiden reduceras och göra klassificeringen mer effektiv. I avhandlingen studeras huruvida av diskretisering påverkar klassificeringens prediceringsnoggrannhet och en mycket effektiv diskretiseringsmetod för högdimensionella data föreslås. Linjära klassificeringsmetoder har fördelen att vara stabila. Nackdelen är att de kräver en inverterbar kovariansmatris och vilket kovariansmatrisen inte är för högdimensionella data. I avhandlingen föreslås ett sätt att skatta inversen för glesa kovariansmatriser med blockdiagonalmatris. Denna matris har dessutom fördelen att det leder till additiv klassificering vilket möjliggör att välja hela block av relevanta variabler. I avhandlingen presenteras även en metod för att identifiera och välja ut blocken. Det finns också probabilistiska klassificeringsmetoder som har fördelen att ge sannolikheten att tillhöra vardera av de möjliga utfallen för en observation, inte som de flesta andra klassificeringsmetoder som bara predicerar utfallet. I avhandlingen förslås en sådan Bayesiansk metod, givet den blockdiagonala matrisen och normalfördelade utfallsklasser. De i avhandlingen förslagna metodernas relevans och fördelar är visade genom att tillämpa dem på simulerade och riktiga högdimensionella data.

The aim of this study is to compare remote sensing methods in the context of a vegetation mapping of an urban environment. The methods used was (1) a traditional per-pixel based method; maximum likelihood supervised classification (ENVI), (2) a standard object based method; example based feature extraction (ENVI) and (3) a newly developed method; Window Independent Contextual Segmentation (WICS) (Choros Cognition). A four-band SPOT5 image with a pixel size of 10x10m was used for the classifications. A validation data-set was created using a ortho corrected aerial image with a pixel size of 1x1m. Error matrices was created by cross-tabulating the classified images with the validation data-set. From the error matrices, overall accuracy and kappa coefficient was calculated. The object-based method performed best with a overall accuracy of 80% and a kappa value of 0.6, followed by the WICS method with an overall accuracy of 77% and a kappa value of 0.53, placing the supervised classification last with an overall accuracy of 71% and a kappa value of 0.38. The results of this study suggests object-based method and WICS to perform better than the supervised classification in an urban environment.

I will investigate applications of machine learning algorithms to medical data, adaptations of differences in data collection, and the use of ensemble techniques. Focusing on the binary classification problem of Parkinson’s Disease (PD) diagnosis, I will apply machine learning algorithms to a primary dataset consisting of voice recordings from healthy and PD subjects. Specifically, I will use Artificial Neural Networks, Support Vector Machines, and an Ensemble Learning algorithm to reproduce results from [MS12] and [GM09]. Next, I will adapt a secondary regression dataset of PD recordings and combine it with the primary binary classification dataset, testing various techniques to consolidate the data including treating the regression data as unlabeled data in a semi-supervised learning approach. I will determine the performance of the above algorithms on this consolidated dataset. Performance of algorithms will be evaluated using 10-fold cross validation and results will be analyzed in a confusion matrix. Accuracy, precision, recall, and F-score will be calculated. The expands on past related work, which has used either a regression dataset alone to predict a Unified Parkinson’s Disease Rating Scale score for PD patients, or a classification dataset to determine healthy or PD diagnosis. In past work, the datasets have not been combined, and the regression set has not been used to contribute to evaluation of healthy subjects.

In this study, an urban growth model is used to simulate the urban growth in 2025 in the Antalya Metropolitan Area. It is the fastest growing metropolis in Turkey with a population growth of 41,79&amp / #8240 / , although Turkey&amp / #8217 / s growth is 18,28&amp / #8240 / for the last decade. An Urban Growth Model (SLEUTH, Version 3.0) is calibrated with cartographic data. The prediction is based on the archived data trends of the years of the 1987, 1996, and 2002 images, which are extracted from Landsat Thematic Mapper and Enhanced Thematic Mapper satellite images and the aerial photographs acquired in 1992 and the data are prepared to insert them as input into the model. The urban extent is obtained through supervised classification of the satellite images and visual interpretation of aerial photographs. The model calibration, where a predetermined order of stepping through the coefficient space is used is performed in order to determine the best fit values for the five growth control parameters including the coefficients of diffusion, breed and spread, slope and road gravity with the historical urban extent data. The development trend in Antalya is simulated by slowing down growth by taking into consideration the road development and environmental protection. After the simulation for a period of 23 years, 9824 ha increased in urban areas is obtained for 2025.

Missense mutations account for more than 50% of the mutations known to be involved in human inherited diseases. Missense classification is a challenging task that involves sequencing of the genome, identifying the variations, and assessing their deleteriousness. This is a very laborious, time and cost intensive task to be carried out in the laboratory. Advancements in bioinformatics have led to several large-scale next-generation genome sequencing projects, and subsequently the identification of genome variations. Several studies have combined this data with information on established deleterious and neutral variants to develop machine learning based classifiers. There are significant issues with the missense classifiers due to which missense classification is still an open area of research. These issues can be classified under two broad categories: (a) Dataset overlap issue - where the performance estimates reported by the state-of-the-art classifiers are overly optimistic as they have often been evaluated on datasets that have significant overlaps with their training datasets. Also, there is no comparative analysis of these tools using a common benchmark dataset that contains no overlap with the training datasets, therefore making it impossible to identify the best classifier among them. Also, such a common benchmark dataset is not available. (b) Inadequate capture of vital biological information of the protein and mutations - such as conservation of long-range amino acid dependencies, changes in certain physico-chemical properties of the wild-type and mutant amino acids, due to the mutation. It is also not clear how to extract and use this information. Also, some classifiers use structural information that is not available for all proteins. In this study, we compiled a new dataset, containing around 2 - 15% overlap with the popularly used training datasets, with 18,036 mutations in 5,642 proteins. We reviewed and evaluated 15 state-of-the-art missense classifiers - SIFT, PANTHER, PROVEAN, PhD-SNP, Mutation Assessor, FATHMM, SNPs&GO, SNPs&GO3D, nsSNPAnalyzer, PolyPhen-2, SNAP, MutPred, PON-P2, CONDEL and MetaSNP, using the six metrics - accuracy, sensitivity, specificity, precision, NPV and MCC. When evaluated on our dataset, we observe huge performance drops from what has been claimed. Average drop in the performance for these 13 classifiers are around 15% in accuracy, 17% in sensitivity, 14% in specificity, 7% in NPV, 24% in precision and 30% in MCC. With this we show that the performance of these tools is not consistent on different datasets, and thus not reliable for practical use in a clinical setting. As we observed that the performance of the existing classifiers is poor in general, we tried to develop a new classifier that is robust and performs consistently across datasets, and better than the state-of-the-art classifiers. We developed a novel method of capturing long-range amino acid dependency conservation by boosting the conservation frequencies of substrings of amino acids of various lengths around the mutation position using AdaBoost learning algorithm. This score alone performed equivalently to the sequence conservation based tools in classifying missense mutations. Popularly used sequence conservation properties was combined with this boosted long-range dependency conservation scores using AdaBoost algorithm. This reduced the class bias, and improved the overall accuracy of the classifier. We trained a third classifier by incorporating changes in 21 important physico-chemical properties, due to the mutation. In this case, we observed that the overall performance further improved and the class bias further reduced. The performance of our final classifier is comparable with the state-of-the-art classifiers. We did not find any significant improvement, but the class-specific accuracies and precisions are marginally better by around 1-2% than those of the existing classifiers. In order to understand our classifier better, we dissected our benchmark dataset into: (a) seen and unseen proteins, and (b) pure and mixed proteins, and analysed the performance in detail. Finally we concluded that our classifier performs consistently across each of these categories of seen, unseen, pure and mixed protein.

Made available in DSpace on 2015-03-26T13:53:32Z (GMT). No. of bitstreams: 1 texto completo.pdf: 3797598 bytes, checksum: 93618b9d18c5fd42a738d4d3d1f869f6 (MD5) Previous issue date: 2013-03-26 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The need for recognition of the vast territory of the São Francisco valley for defining priority areas for the implementation of recovery actions and environmental preservation motivated the proposition of this work. We attempted to find low-cost methods for performing this task on submedium, which includes drainage shedding for the trough between the municipalities of Bahia Pilão Arcado and Paulo Afonso. The sub-basin of the river Pacuí was taken as a pilot area for this job. Located in northern Bahia, has 1010,1 km2 and is fully inserted in the municipality of Campo Formoso. It is a tributary of the Salitre River, which rises in the Chapada Diamantina north and empties into the São Francisco river in Juazeiro, 20 km upstream of the seat. The sub-basin has just over 10.000 inhabitants and the main town is existing is Lajes dos Negros, which is distant 96 km from the municipal seat. The region has an average annual temperature of 23,7 oC, evapotranspiration from 1000 to 1400 mm, annual rainfall between 475 and 700 mm, according to the influence of relief, characterized as semi-arid climate , the type classification BSh by Koppen-Geiger. The mountains that surround the dividers topographic northern and western reaches maximum elevation of 1275 m. The center, south and east of the basin have gentle slope and elevation of the mouth is 465 m. Was tested using TM / Landsat 5 to find the best combinations of bands for the distinction of classes of use and land use in the lower basin San Francisco through the Maximum Likelihood classifier. We used the natural sensor TM bands except 6, which deals with the thermal infrared. Were produced four bands to participate in the artificial combinations: the first three bands of the principal component analysis (PCA) and vegetation index (NDVI). We analyzed 1023 different combinations. It was noticed that the PCA3 adds no good results due to the low amount of information. The band also has 2 bad performance often, but there are some cases where benefits. As for the other, it is ideal to use all or at best remove some of the visible region. The combination of natural and artificial bands is ideal. NDVI, PCA1, bands 4, 5 and 7 are the best that deliver results. With the selected bands, recognized the evolution of land use in the sub-basin of the river Pacuí through images TM/Lansat-5 current and historical. The result was the existence of more than 28.000 ha of heavily disturbed areas in the sub-basin, with 45% of the valley and 8% of the mountains are covered with vegetation degraded. We also used the NDVIs four TM/Landsat-5 to investigate deforestation and regeneration of vegetation cover in the last two decades. We employed a subtraction method of vegetation indices. The images used were from August 1992, May 2001, October 2001 and June 2011. For the production of NDVIs were made radiometric calibration, generation of surface reflectance and standardization of means and variances. The subtraction images were divided according to the twelve classes defined for use and occupation for the oldest image of the pair comparison. Each sector was analyzed fitting image pixels difference under nine different ranges: high, medium, low and very low deforestation; remain unchanged, and very low, low, medium and high regeneration. The intervals were calculated from the average by summing up one or down one and a half, two or three times the standard deviation respectively. Were dismissed as clearing those areas that were once agricultural soils or exposed. Only regeneration areas that were considered in more recent image were not classified as agriculture and in the past were exposed soils or pastures. The result was over 1000 ha of accumulated deforestation from 1992 to 2011. We also analyzed the characteristics of the soils of sub-basin through eleven soil profiles were determined and 10 different environments occurrence. Performed analyzes were physical, chemical and morphological routine. We have found the following classes: Neossolo Litólico Eutrófico típico, Cambissolo Háplico Tb Distrófico latossólico, Latossolo Vermelho- Amarelo Distrófico típico (2 profiles), Neossolo Litólico Distrófico típico, Neossolo Quartzarênico Hidromórfico típico, Cambissolo Háplico Tb Eutrófico latossólico (2 profiles), Cambissolo Háplico Tb Eutrófico léptico, Cambissolo Háplico Carbonático saprolítico e Neossolo Quartzarênico Órtico latossólico. It was noticed that the relief is a major factor for the differentiation of environments occurring in the more hilly the Neossolos Litólicos, the intermediate parts Cambissolos and parts flatter Latossolos. Soils that occur on limestone result eutrophic and on sandstone and detrital cover, dystrophic. The portions of the landscape under influence of Formation Bebedouro have soils rich in silt, strongly alkaline and high susceptibility to erosion. Any areas within the basin where it promotes loss of permeability shows strong signs of erosion. The mountainous region of the sub-basin and some areas of sandy soils for lowland present good condition of vegetation and constitute one of the last preserved parts of the region, and may account for much of the recharge of underground springs that keeps the river Pacuí perennial. Should be preserved, along with the huge speleological site. It is necessary to change the means of production based on unbridled exploitation of natural resources to activities that approximate environmental sustainability. The activity most striking is the extensive goat and sheep. It must be recovered to cover the slopes of the river Pacuí, the main focus of erosion. It should also control the runoff of roads, settlements and vegetation cover to shedding impoverished areas of greatest slope. / A necessidade de reconhecimento do vasto território da bacia do rio São Francisco para a definição de áreas prioritárias para a aplicação de ações de recuperação e preservação ambiental motivou a proposição deste trabalho.Buscou-se encontrar métodos de baixo custo para a execução desta tarefa no trecho submédio, que engloba a drenagem que verte para a calha entre os municípios baianos de Pilão Arcado e Paulo Afonso. A sub-bacia do rio Pacuí foi tomada como área piloto para este trabalho. Localiza-se no norte da Bahia, possui 1010,1 km² e está totalmente inserida no município de Campo Formoso. É afluente do rio Salitre, o qual nasce na Chapada Diamantina norte e deságua no rio São Francisco em Juazeiro, 20 km a montante da sede. A sub-bacia tem pouco mais de 10.000 habitantes e o principal povoado existente é Laje dos Negros, que dista 96 km da sede municipal. A região possui temperatura média anual de 23,7 ºC, evapotranspiração de 1000 a 1400 mm e precipitação anual entre 475 e 700 mm, de acordo com a influência do relevo, caracterizando o clima como semiárido, do tipo BSh pela classificação de Koppen-Geiger. As serras que delimitam os divisores topográficos norte e oeste atingem cota máxima de 1275 m. O centro, sul e leste da bacia possuem relevo suave e a cota da foz é 465 m. Foi testado o uso de imagens TM/Landsat 5 para encontrar as melhores combinações de bandas para a distinção de classes de uso e ocupação dos solos no Submédio São Francisco por meio do classificador Máxima Verossimilhança. Foram utilizadas as bandas naturais do sensor TM, exceto a 6, que trata do infravermelho termal. Foram produzidas quatro bandas artificiais para participar das combinações: as três primeiras bandas da análise de componentes principais (PCAs) e o índice de vegetação por diferença normalizada (NDVI). Foram analisadas 1023 diferentes combinações. Percebeu-se que a PCA3 não agrega bons resultados devido à sua baixa quantidade de informação. A banda 2 também tem mal desempenho frequentemente, mas há alguns casos em que traz benefícios. Quanto às demais, o ideal é utilizar todas ou no máximo remover algumas da região do visível. A combinação entre as bandas naturais e artificiais é o ideal. NDVI, PCA1, bandas 4, 5 e 7 são as que agregam melhores resultados. Com as bandas escolhidas, foi reconhecida a evolução do uso do solo na sub-bacia do rio Pacuí por meio de imagens TM/Lansat-5 atuais e históricas. O resultado foi a existência de mais de 28.000 ha de áreas fortemente antropizadas na sub-bacia, sendo que 45 % do vale e 8 % das serras estão com a cobertura vegetal degradada. Foram utilizados ainda os NDVIs de quatro imagens TM/Landsat-5 para investigar desmatamento e regeneração da cobertura vegetal nas duas últimas décadas. Foi empregado o método de subtração dos índices de vegetação. As imagens utilizadas foram de agosto de 1992, maio de 2001, outubro de 2001 e junho de 2011. Para a produção dos NDVIs foram feitas calibração radiométrica, geração de reflectância de superfície e uniformização de médias e variâncias. As imagens de subtração foram divididas segundo as doze classes de uso e ocupação definidas para a imagem mais antiga do par de comparação. Cada setor foi analisado enquadrando pixels de imagens diferença segundo nove diferentes intervalos: alto, médio, baixo e muito baixo desmatamento; inalteração; e muito baixa, baixa, média e alta regeneração. Os intervalos foram calculados a partir da média somando-se ou diminuindo-se uma, uma e meia, duas ou três vezes o desvio padrão respectivamente. Foram desconsideradas como desmatamento aquelas áreas que no passado eram agricultura ou solos expostos. Só foram consideradas regeneração áreas que na imagem mais recente não foram classificadas como agricultura e na do passado eram pastagens ou solos expostos. O resultado foi mais de 1000 ha de desmatamento acumulado no período 1992 a 2011. Analisou-se também as características dos solos da sub-bacia por meio de onze perfís de solo e foram determinados 10 diferentes ambientes de ocorrência. Realizaram-se análises físicas, químicas e morfológicas de rotina. Foram encontradas as seguintes classes: Neossolo Litólico Eutrófico típico, Cambissolo Háplico Tb Distrófico latossólico, Latossolo Vermelho-Amarelo Distrófico típico (2 perfis), Neossolo Litólico Distrófico típico, Neossolo Quartzarênico Hidromórfico típico, Cambissolo Háplico Tb Eutrófico latossólico (2 perfís), Cambissolo Háplico Tb Eutrófico léptico, Cambissolo Háplico Carbonático saprolítico e Neossolo Quartzarênico Órtico latossólico. Percebeu-se que o relevo é um dos fatores principais para a diferenciação dos pedoambientes, ocorrendo nas partes mais declivosas os Neossolos litólicos, nas partes intermediárias Cambissolos e nas partes mais planas Latossolos. Solos que ocorrem sobre calcário resultam eutróficos e sobre arenito e cobertura detríticas, distróficos. As porções da paisagem sob influencia da Formação Bebedouro possuem solos ricos em silte, fortemente alcalinos e com alta suscetibilidade à erosão. Quaisquer áreas dentro da bacia onde se promova perda da permeabilidade apresenta fortes sinais de erosão. A região serrana da sub-bacia e algumas áreas de solos arenosos de baixada apresentam bom estado de conservação da cobertura vegetal e constituem uma das últimas partes preservadas da região, e podem responder por grande parte da recarga dos mananciais subterrâneos que mantém o rio Pacuí perene. Devem ser preservadas, juntamente com o enorme patrimônio espeleológico local. Se faz necessária a mudança dos meios de produção baseados na exploração desregrada dos recursos naturais, para atividades que se aproximem da sustentabilidade ambiental. A atividade mais impactante é a caprino-ovinocultura extensiva. É preciso que seja recuperada a cobertura vegetal das encostas do rio Pacuí, principal foco de processos erosivos. Deve-se ainda controlar o escoamento superficial em estradas, povoados e em áreas com cobertura vegetal empobrecida que vertem para locais de maior declividade.

In this dissertation, an ensemble non-uniform spectral feature selection and a kernel density decision fusion framework are proposed for the classification of hyperspectral data using a support vector machine classifier. Hyperspectral data has more number of bands and they are always highly correlated. To utilize the complete potential, a feature selection step is necessary. In an ensemble situation, there are mainly two challenges: (1) Creating diverse set of classifiers in order to achieve a higher classification accuracy when compared to a single classifier. This can either be achieved by having different classifiers or by having different subsets of features for each classifier in the ensemble. (2) Designing a robust decision fusion stage to fully utilize the decision produced by individual classifiers. This dissertation tests the efficacy of the proposed approach to classify hyperspectral data from different applications. Since these datasets have a small number of training samples with larger number of highly correlated features, conventional feature selection approaches such as random feature selection cannot utilize the variability in the correlation level between bands to achieve diverse subsets for classification. In contrast, the approach proposed in this dissertation utilizes the variability in the correlation between bands by dividing the spectrum into groups and selecting bands from each group according to its size. The intelligent decision fusion proposed in this approach uses the probability density of training classes to produce a final class label. The experimental results demonstrate the validity of the proposed framework that results in improvements in the overall, user, and producer accuracies compared to other state-of-the-art techniques. The experiments demonstrate the ability of the proposed approach to produce more diverse feature selection over conventional approaches.