This thesis addresses three research gaps in published literature. These are, the absence of Object Based Image Analysis (OBIA) methods for urban Land Use and Land Cover (LULC) analysis in Nigeria; the inability to use Nigeriasat-1 satellite data for urban LULC analysis and monitoring urban growth in Nigeria with Shannon’s Entropy Index. Using Abuja as a case study, this research investigated the nature of land use/land cover change (LULCC). Specific objectives were: design of an object based classification method to extract urban LULC; validate a method to extract LULC in developing countries from multiple sources of remotely sensed data; apply the method to extract LULC data; use the outputs to validate an Urban Growth Model (UGM); optimise an UGM to represent patterns and trends and through this iterative process identify and prioritise the driving forces of urban change; and finally use the outputs of the land use maps to determine if planning has controlled land use development. Landsat 7 ETM (2001), Nigeriasat-1 SLIM (2003) and SPOT 5 HRG (2006) sensor data were merged with land use cadastre in OBIA, to produce land use maps. Overall classification accuracies were 92%, 89% and 96% respectively. Post classification analysis of LULCC indicated 4.43% annual urban spread. Shannon’s Entropy index for the study period were 0.804 (2001), 0.898 (2003) and 0.930 (2006). Cellular Automata/Markov analysis was also used to predict urban growth trend of 0.89% per annum. For the first time OBIA has been used for LULC analysis in Nigeria. This research has established that Nigeriasat-1 data can contribute to urban studies using innovative OBIA methods. In addition, that Shannon’s Entropy Index can be used to understand the nature of urban growth in Nigeria. Finally, the drivers of LULCC in Abuja are similar to those of planned capital cities in other developing economies. Land use developments in Abuja can provide an insight into urban dynamics in a developing country’s capital region. OBIA, Shannon’s Entropy Index and UGM can aid urban administrators and provide information for sustainable urban planning and development.

Forest ecosystems have recently received worldwide attention due to their biological diversity and their major role in the global carbon balance. Detecting forest cover change is crucial for reporting forest status and assessing the evolution of forested areas. However, existing change detection approaches based on satellite remote sensing are not quite appropriate to rapidly process the large volume of earth observation data. Recent advances in image segmentation have led to new opportunities for a new object-based monitoring system. <br> <br> This thesis aims at developing and evaluating an automated object-based change detection method dedicated to high spatial resolution satellite images for identifying and mapping forest cover changes in different ecosystems. This research characterized the spectral reflectance dynamics of temperate forest stand cycle and found the use of several spectral bands better for the detection of forest cover changes than with any single band or vegetation index over different time periods. Combining multi-date image segmentation, image differencing and a dedicated statistical procedure of multivariate iterative trimming, an automated change detection algorithm was developed. This process has been further generalized in order to automatically derive an up-to-date forest mask and detect various deforestation patterns in tropical environment.<br> <br> Forest cover changes were detected with very high performances (>90 %) using 3 SPOT-HRVIR images over temperate forests. Furthermore, the overall results were better than for a pixel-based method. Overall accuracies ranging from 79 to 87% were achieved using SPOT-HRVIR and Landsat ETM imagery for identifying deforestation for two different case studies in the Virunga National Park (DRCongo). Last but not least, a new multi-scale mapping solution has been designed to represent change processes using spatially-explicit maps, i.e. deforestation rate maps. By successfully applying these complementary conceptual developments, a significant step has been done toward an operational system for monitoring forest in various ecosystems.

Understanding, monitoring and managing savanna ecosystems require characterizing both functional and structural properties of vegetation. Due to functional diversity and structural heterogeneity in savannas, characterizing these properties using remote sensing is methodologically challenging. Focusing on the semi-arid savanna in the central Kalahari, the objective of this dissertation was to combine in situ data with multi-scale satellite imagery and two image analysis approaches (i.e. Multiple Endmember Spectral Mixture Analysis (MESMA) and Object Based Image Analysis (OBIA)) to : (i) determine the superior method for estimating fractional photosynthetic vegetation (fPV), non-photosynthetic vegetation (fNPV) and bare soil (fBS) when high spatial resolution multispectral imagery is used, (ii) examine the suitability of OBIA for mapping vegetation morphology types using a Landsat TM imagery, (iii) examine the impact of changing spatial resolution on magnitude and accuracy of fractional cover and (iv) examine how the fractional cover magnitude and accuracy are spatially associated with vegetation morphology. Using the GeoEye-1 imagery, MESMA provided more accurate fractional cover estimates than OBIA. The increasing segmentation scale in OBIA resulted in a consistent increase in error. While areas under woody cover produced lower errors even at coarse segmentation scales, those with herbaceous cover provided low errors only at the fine segmentation scale. Vegetation morphology type mapping results suggest that classes with dominant woody life forms attained higher accuracy at fine segmentation scales, while those with dominant herbaceous vegetation reached higher classification accuracy at coarse segmentation scales. Contrarily, for bare areas accuracy was relatively unaffected by changing segmentation scale. Multi-scale fractional cover mapping results indicate that increasing pixel size caused consistent increases in variance of and error in fractional cover estimates. Even at a coarse spatial resolution, fPV was estimated with higher accuracy compared to fNPV and fBS. At a larger pixel size, in areas with dominant woody vegetation, fPV was overestimated at the cost of mainly underestimating fBS; in contrast, in areas with dominant herbaceous vegetation, fNPV was overestimated with a corresponding underestimation of both fPV and fBS. These results underscore that structural and functional heterogeneity in savannas impact retrieval of fractional cover, suggesting that comprehensive remote sensing of savannas needs to take both structure and cover into account. / text

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.

Câmaras não métricas acopladas a Veículos Aéreos Não Tripulados (VANT) possibilitam coleta de imagens com alta resolução espacial e temporal. Além disso, o custo de operação e manutenção desses equipamentos são reduzidos. A classificação da cobertura da terra por meio dessas imagens são dificultadas devido à alta variabilidade espectral dos alvos e ao grande volume de dados gerados. Esses contratempos são contornados utilizando Análise de Imagens Baseada em Objetos (Object-Based Image Analysis – OBIA) e algoritmos de mineração de dados. Um algoritmo empregado na OBIA são as Árvores de Decisão (AD). Essa técnica possibilita tanto a seleção de atributos mais informativos quanto a classificação das regiões. Novas técnicas de AD foram desenvolvidas e, nessas inovações, foram inseridas funções para selecionar atributos e para melhorar a classificação. Um exemplo é o algoritmo C5.0, que possui uma função de redução de dados e uma de reforço. Nesse contexto, este trabalho tem como objetivo (i) avaliar o método de segmentação por crescimento de regiões em imagens com altíssima resolução espacial, (ii) determinar os atributos preditivos mais importantes na discriminação das classes e (iii) avaliar as classificações das regiões em relação aos parâmetros de seleção dos atributos (winnow) e de reforço (trial), que estão contidos no algoritmo C5.0. A segmentação da imagem foi efetuada no programa Spring, já as regiões geradas na segmentação foram classificadas pelo modelo de AD C5.0, que está disponível no programa R. Como resultado foi identificado que a segmentação crescimento de regiões possibilitou uma alta correspondência com regiões geradas pelo especialista, resultando em valores de Reference Bounded Segments Booster (RBSB) próximos a 0. Os atributos mais importantes na construção dos modelos por AD foram a razão entre a banda do verde com a azul (r_v_a) e o Modelo Digital de Elevação (MDE). Para o parâmetro de reforço (trial), não foi identificada melhora na acurácia da classificação ao aumentar seu valor. Já o parâmetro winnow possibilitou uma redução no número de atributos preditivos, sem perdas estatisticamente significativas na acurácia da classificação. A função de reforço (trial) não melhorou a classificação da cobertura da terra. Também não foram constatadas diferenças estatisticamente significativas quando winnow selecionado como verdadeiro, mas se encontrou o benefício desse último parâmetro reduzindo a dimensionalidade dos dados. Nesse sentido, este trabalho contribuiu para a classificação da cobertura da terra em imagens coletadas por VANT, uma vez que se desenvolveu algoritmos para automatizar os processos da OBIA e para avaliar a classificação das regiões em relação às funções de reforço (winnow) e de seleção do atributo (winnow) do classificador por árvore de decisão C5.0. / Non-metric cameras attached to Unmanned Aerial Vehicles (UAV) enable collection of images with high spatial and temporal resolution. In addition, the cost of operation and maintenance of equipment are reduced. The land cover classification through these images are hampered due to high spectral variability of the targets and the large volume of data generated. These setbacks are contoured using Image Analysis Based on Objects (OBIA) and data mining algorithms. An algorithm used in OBIA are Decision Trees (AD). This technique allows the selection of the most informative attributes as the classification of regions. New AD techniques have been developed and these innovations, were functions inserted to select attributes and to improve classification. One example is a C5.0 algorithm, which has a data reduction function and of boosting. In this context, this paper aims to (i) evaluate the segmentation method for growing regions in images with high spatial resolution, (ii) determine the most important predictive attributes in the discrimination of classes and (iii) evaluate the classifications of regions regarding the attributes selection parameters (winnow) and boosting (trial), which are contained in the C5.0 algorithm. The image segmentation was performed in Spring program, since the regions generated in segmentation were classified by model C5.0 , which is available in the program R. As a result it was identified that the segmentation by region growing provided a high correlation with regions generated by the expert, resulting in Reference Bounded Segments Booster values (RBSB) near 0. The most important features in the construction of models of decision tree are the ratio between the band of green with the blue (r_v_a) and the Digital Elevation Model (DEM). Was not identified improvement in classification accuracy when was increased value of trial parameter. Already winnow parameter enabled a reduction in the number of predictive attributes, with no statistically significant losses in the accuracy of the classification. The boosting function (trial) did not improve the classification of land cover. Also were not found statistically significant differences when winnow selected as true, but was found the benefit of the latter parameter to reducing the dimensionality of the data. Thus, this work contributed to the land cover classification in images collected by UAV, once that were developed algorithms to automate the processes of integration OBIA and decision tree (C5.0).

Câmaras não métricas acopladas a Veículos Aéreos Não Tripulados (VANT) possibilitam coleta de imagens com alta resolução espacial e temporal. Além disso, o custo de operação e manutenção desses equipamentos são reduzidos. A classificação da cobertura da terra por meio dessas imagens são dificultadas devido à alta variabilidade espectral dos alvos e ao grande volume de dados gerados. Esses contratempos são contornados utilizando Análise de Imagens Baseada em Objetos (Object-Based Image Analysis – OBIA) e algoritmos de mineração de dados. Um algoritmo empregado na OBIA são as Árvores de Decisão (AD). Essa técnica possibilita tanto a seleção de atributos mais informativos quanto a classificação das regiões. Novas técnicas de AD foram desenvolvidas e, nessas inovações, foram inseridas funções para selecionar atributos e para melhorar a classificação. Um exemplo é o algoritmo C5.0, que possui uma função de redução de dados e uma de reforço. Nesse contexto, este trabalho tem como objetivo (i) avaliar o método de segmentação por crescimento de regiões em imagens com altíssima resolução espacial, (ii) determinar os atributos preditivos mais importantes na discriminação das classes e (iii) avaliar as classificações das regiões em relação aos parâmetros de seleção dos atributos (winnow) e de reforço (trial), que estão contidos no algoritmo C5.0. A segmentação da imagem foi efetuada no programa Spring, já as regiões geradas na segmentação foram classificadas pelo modelo de AD C5.0, que está disponível no programa R. Como resultado foi identificado que a segmentação crescimento de regiões possibilitou uma alta correspondência com regiões geradas pelo especialista, resultando em valores de Reference Bounded Segments Booster (RBSB) próximos a 0. Os atributos mais importantes na construção dos modelos por AD foram a razão entre a banda do verde com a azul (r_v_a) e o Modelo Digital de Elevação (MDE). Para o parâmetro de reforço (trial), não foi identificada melhora na acurácia da classificação ao aumentar seu valor. Já o parâmetro winnow possibilitou uma redução no número de atributos preditivos, sem perdas estatisticamente significativas na acurácia da classificação. A função de reforço (trial) não melhorou a classificação da cobertura da terra. Também não foram constatadas diferenças estatisticamente significativas quando winnow selecionado como verdadeiro, mas se encontrou o benefício desse último parâmetro reduzindo a dimensionalidade dos dados. Nesse sentido, este trabalho contribuiu para a classificação da cobertura da terra em imagens coletadas por VANT, uma vez que se desenvolveu algoritmos para automatizar os processos da OBIA e para avaliar a classificação das regiões em relação às funções de reforço (winnow) e de seleção do atributo (winnow) do classificador por árvore de decisão C5.0. / Non-metric cameras attached to Unmanned Aerial Vehicles (UAV) enable collection of images with high spatial and temporal resolution. In addition, the cost of operation and maintenance of equipment are reduced. The land cover classification through these images are hampered due to high spectral variability of the targets and the large volume of data generated. These setbacks are contoured using Image Analysis Based on Objects (OBIA) and data mining algorithms. An algorithm used in OBIA are Decision Trees (AD). This technique allows the selection of the most informative attributes as the classification of regions. New AD techniques have been developed and these innovations, were functions inserted to select attributes and to improve classification. One example is a C5.0 algorithm, which has a data reduction function and of boosting. In this context, this paper aims to (i) evaluate the segmentation method for growing regions in images with high spatial resolution, (ii) determine the most important predictive attributes in the discrimination of classes and (iii) evaluate the classifications of regions regarding the attributes selection parameters (winnow) and boosting (trial), which are contained in the C5.0 algorithm. The image segmentation was performed in Spring program, since the regions generated in segmentation were classified by model C5.0 , which is available in the program R. As a result it was identified that the segmentation by region growing provided a high correlation with regions generated by the expert, resulting in Reference Bounded Segments Booster values (RBSB) near 0. The most important features in the construction of models of decision tree are the ratio between the band of green with the blue (r_v_a) and the Digital Elevation Model (DEM). Was not identified improvement in classification accuracy when was increased value of trial parameter. Already winnow parameter enabled a reduction in the number of predictive attributes, with no statistically significant losses in the accuracy of the classification. The boosting function (trial) did not improve the classification of land cover. Also were not found statistically significant differences when winnow selected as true, but was found the benefit of the latter parameter to reducing the dimensionality of the data. Thus, this work contributed to the land cover classification in images collected by UAV, once that were developed algorithms to automate the processes of integration OBIA and decision tree (C5.0).

Câmaras não métricas acopladas a Veículos Aéreos Não Tripulados (VANT) possibilitam coleta de imagens com alta resolução espacial e temporal. Além disso, o custo de operação e manutenção desses equipamentos são reduzidos. A classificação da cobertura da terra por meio dessas imagens são dificultadas devido à alta variabilidade espectral dos alvos e ao grande volume de dados gerados. Esses contratempos são contornados utilizando Análise de Imagens Baseada em Objetos (Object-Based Image Analysis – OBIA) e algoritmos de mineração de dados. Um algoritmo empregado na OBIA são as Árvores de Decisão (AD). Essa técnica possibilita tanto a seleção de atributos mais informativos quanto a classificação das regiões. Novas técnicas de AD foram desenvolvidas e, nessas inovações, foram inseridas funções para selecionar atributos e para melhorar a classificação. Um exemplo é o algoritmo C5.0, que possui uma função de redução de dados e uma de reforço. Nesse contexto, este trabalho tem como objetivo (i) avaliar o método de segmentação por crescimento de regiões em imagens com altíssima resolução espacial, (ii) determinar os atributos preditivos mais importantes na discriminação das classes e (iii) avaliar as classificações das regiões em relação aos parâmetros de seleção dos atributos (winnow) e de reforço (trial), que estão contidos no algoritmo C5.0. A segmentação da imagem foi efetuada no programa Spring, já as regiões geradas na segmentação foram classificadas pelo modelo de AD C5.0, que está disponível no programa R. Como resultado foi identificado que a segmentação crescimento de regiões possibilitou uma alta correspondência com regiões geradas pelo especialista, resultando em valores de Reference Bounded Segments Booster (RBSB) próximos a 0. Os atributos mais importantes na construção dos modelos por AD foram a razão entre a banda do verde com a azul (r_v_a) e o Modelo Digital de Elevação (MDE). Para o parâmetro de reforço (trial), não foi identificada melhora na acurácia da classificação ao aumentar seu valor. Já o parâmetro winnow possibilitou uma redução no número de atributos preditivos, sem perdas estatisticamente significativas na acurácia da classificação. A função de reforço (trial) não melhorou a classificação da cobertura da terra. Também não foram constatadas diferenças estatisticamente significativas quando winnow selecionado como verdadeiro, mas se encontrou o benefício desse último parâmetro reduzindo a dimensionalidade dos dados. Nesse sentido, este trabalho contribuiu para a classificação da cobertura da terra em imagens coletadas por VANT, uma vez que se desenvolveu algoritmos para automatizar os processos da OBIA e para avaliar a classificação das regiões em relação às funções de reforço (winnow) e de seleção do atributo (winnow) do classificador por árvore de decisão C5.0. / Non-metric cameras attached to Unmanned Aerial Vehicles (UAV) enable collection of images with high spatial and temporal resolution. In addition, the cost of operation and maintenance of equipment are reduced. The land cover classification through these images are hampered due to high spectral variability of the targets and the large volume of data generated. These setbacks are contoured using Image Analysis Based on Objects (OBIA) and data mining algorithms. An algorithm used in OBIA are Decision Trees (AD). This technique allows the selection of the most informative attributes as the classification of regions. New AD techniques have been developed and these innovations, were functions inserted to select attributes and to improve classification. One example is a C5.0 algorithm, which has a data reduction function and of boosting. In this context, this paper aims to (i) evaluate the segmentation method for growing regions in images with high spatial resolution, (ii) determine the most important predictive attributes in the discrimination of classes and (iii) evaluate the classifications of regions regarding the attributes selection parameters (winnow) and boosting (trial), which are contained in the C5.0 algorithm. The image segmentation was performed in Spring program, since the regions generated in segmentation were classified by model C5.0 , which is available in the program R. As a result it was identified that the segmentation by region growing provided a high correlation with regions generated by the expert, resulting in Reference Bounded Segments Booster values (RBSB) near 0. The most important features in the construction of models of decision tree are the ratio between the band of green with the blue (r_v_a) and the Digital Elevation Model (DEM). Was not identified improvement in classification accuracy when was increased value of trial parameter. Already winnow parameter enabled a reduction in the number of predictive attributes, with no statistically significant losses in the accuracy of the classification. The boosting function (trial) did not improve the classification of land cover. Also were not found statistically significant differences when winnow selected as true, but was found the benefit of the latter parameter to reducing the dimensionality of the data. Thus, this work contributed to the land cover classification in images collected by UAV, once that were developed algorithms to automate the processes of integration OBIA and decision tree (C5.0).

Land cover mapping is an important part of resource management, planning, and economic predictions. Improvements in remote sensing, machine learning, image processing, and object based image analysis (OBIA) has made the process of identifying land cover types increasingly faster and reliable but these advances are unable to utilize the amount of information encompassed within ultra-high (sub-meter) resolution imagery. Previously, users have typically reduced the resolution of imagery in an attempt to more closely represent the interpretation or object scale in an image and rid the image of any extraneous information within the image that may cause the OBIA process to identify too small of objects when performing semi-automated delineation of objects based on an images’ properties (Mas et al., 2015; Eiesank et al., 2014; Hu et al., 2010). There have been few known attempts to try and maximize this detailed information in high resolution imagery using advanced textural components. In this study we try to circumnavigate the inherent problems associated with high resolution imagery by combining well researched data transformations that aid the OBIA process with a seldom used texture transformation in Geographic Object Based Image Analyses (GEOBIA) known as the Gabor Transform and the hierarchal organization of landscapes. We will observe the difference made in segmentation and classification accuracy of a random forest classifier when we fuse a Gabor transformed image to a Normalized Difference Vegetation Index (NDVI), high resolution multi-spectral imagery (RGB and NIR) and Light Detection and Ranging (LiDAR) derived canopy height model (CHM) within a riparian area in Southeast Iowa. Additionally, we will observe the effects on classification accuracy when adding multi-scale land cover data to objects. Both, the addition of hierarchical information and Gabor textural information, could aid the GEOBIA process in delineating and classifying the same objects that human experts would delineate within this riparian landscape.

Object-based image analysis (OBIA) has been increasingly used to identify terrain features of archaeological sites, but only recently to extract subsurface archaeological features from geophysical data. In this study, we use a semi-automated OBIA to identify Archaic (8000-1000 BC) hearths from Ground-Penetrating Radar (GPR) data collected at David Crockett Birthplace State Park in eastern Tennessee in the southeastern United States. The data were preprocessed using GPR-SLICE, Surfer, and Archaeofusion software, and amplitude depth slices were selected that contained anomalies ranging from 0.80 to 1.20 m below surface (BS). Next, the data were segmented within ESRI ArcMap GIS software using a global threshold and, after vectorization, classified using four attributes: area, perimeter, length-to-width ratio, and Circularity Index. The user-defined parameters were based on an excavated Archaic circular hearth found at a depth greater than one meter, which consisted of fire-cracked rock and had a diameter greater than one meter. These observations were in agreement with previous excavations of hearths at the site. Features that had a high probability of being Archaic hearths were further delineated by human interpretation from radargrams and then ground-truthed by auger testing. The semi-automated OBIA successfully predicted 15 probable Archaic hearths at depths ranging from 0.85 to 1.20 m BS. Observable spatial clustering of hearths may indicate episodes of seasonal occupation by small mobile groups during the Archaic Period.

No description available.