Supervised Classification Process Research Articles

Construction and implementation of a poplar spectral library based on phenological stages for land cover classification using high-resolution satellite images

ABSTRACT As valued members of forest ecosystems, fast-growing tree species are of great importance for protecting natural forests, fulfilling the demand for raw wood material, and preserving biodiversity and ecosystem services. In this regard, multi-temporal monitoring and evaluation of poplar planted areas are essential for decision makers and planners to keep the inventory maps of planted areas up-to-date and track the temporal changes accurately. Using a spectral library, available spectral knowledge can be incorporated into inventory mapping and image classification processes. In this study, in situ spectral measurements of five Populus (four clones and one species) planted sites in Türkiye were collected at three phenological stages (early, mid, and late vegetation) and a novel spectral library was created after completing strict processing stages. Analysis of the spectral reflectances of the Populus species revealed that their spectral properties in the visible region are similar, whereas the discrimination in the near-infrared and shortwave infrared (SWIR) regions was noticeably higher. Two types of applications were conducted to effectively use the spectral library. First, the spectral library was used to validate the poplar class labels of the sites visited during the field campaign. The effectiveness of Worldview-3 imagery with eight SWIR bands was evaluated by pairwise comparison of the spectral curves of the sample fields with those of the spectral library. Second, the spectral library facilitated the extraction of poplar samples from Sentinel-2 imagery in an unvisited region, and a supervised classification process was conducted using the random forest algorithm. The results revealed that the Populus species (Samsun and I-214) could be delineated with high accuracy (>70%) using 10-band Sentinel-2 imagery, and object-based classification outperformed pixel-based classification by approximately 5% for the poplar types. The developed spectral library stands as a valuable asset for studies focused on identifying and classifying Populus species using high-resolution imagery.

Object-Based Image Analysis (OBIA) and Machine Learning (ML) Applied to Tropical Forest Mapping Using Sentinel-2

The purpose of this research was to distinguish and estimate natural forest areas at Paraná, Brazil. Forest plantations (Silviculture) and natural forests have high annual vegetative vigor, as well as agricultural areas in the periods of agricultural harvests, which can bring classification errors between these classes of Land Use and Land Cover (LULC), these classes have similar spectral signatures, but have a distinct texture that can be separated in the supervised classification process, with the joining of object and pixel-to-pixel classification method approaches. Thus, image segmentation techniques through Object-Based Image Analysis (OBIA) and Machine Learning (ML) made forest mapping possible over a large territorial extension. The Google Earth Engine (GEE) platform was used to calculate the vegetation indices (VIs) and Spectral Mixture Analysis (SMA) fraction spectral from Sentinel-2 images, and the creation of homogeneous spectrally shaped regions under supervised classification of phytoecological regions and mesoregions. The overall precision obtained in the mappings resulted in 0.94 Kappa Index (KI) and 96% of Overall Accuracy (OA), which indicates a high performance in large-scale forest mapping. The proposed dataset, source codes and trained models are available on Github (https://github.com/Cechim/simepar-brazil/), creating opportunities for further ad vances in the field.

Urban Growth Forecast Using Machine Learning Algorithms and GIS-Based Novel Techniques: A Case Study Focusing on Nasiriyah City, Southern Iraq

Land use and land cover changes driven by urban sprawl has accelerated the degradation of ecosystem services in metropolitan settlements. However, most optimisation techniques do not consider the dynamic effect of urban sprawl on the spatial criteria on which decisions are based. In addition, integrating the current simulation approach with land use optimisation approaches to make a sustainable decision regarding the suitable site encompasses complex processes. Thus, this study aims to innovate a novel technique that can predict urban sprawl for a long time and can be simply integrated with optimisation land use techniques to make suitable decisions. Three main processes were applied in this study: (1) a supervised classification process using random forest (RF), (2) prediction of urban growth using a hybrid method combining an artificial neural network and cellular automata and (3) the development of a novel machine learning (ML) model to predict urban growth boundaries (UGBs). The ML model included linear regression, RF, K-nearest neighbour and AdaBoost. The performance of the novel ML model was effective, according to the validation metrics that were measured by the four ML algorithms. The results show that the Nasiriyah City expansion (the study area) is haphazard and unplanned, resulting in disastrous effects on urban and natural systems. The urban area ratio was increased by about 10%, i.e., from 2.5% in the year 1992 to 12.2% in 2022. In addition, the city will be expanded by 34%, 25% and 19% by the years 2032, 2042 and 2052, respectively. Therefore, this novel technique is recommended for integration with optimisation land use techniques to determine the sites that would be covered by the future city expansion.

Classification of Alteration Zones Based on Drill Core Hyperspectral Data Using Semi-Supervised Adversarial Autoencoder: A Case Study in Pulang Porphyry Copper Deposit, China

With the development of hyperspectral technology, it has become possible to classify alteration zones using hyperspectral data. Since various altered rocks are comprehensive manifestations of mineral assemblages, their spectra are highly similar, which greatly increases the difficulty of distinguishing among them. In this study, a Semi-Supervised Adversarial Autoencoder (SSAAE) was proposed to classify the alteration zones, using the drill core hyperspectral data collected from the Pulang porphyry copper deposit. The multiscale feature extractor was first integrated into the encoder to fully exploit and mine the latent feature representations of hyperspectral data, which were further transformed into discrete class vectors using a classifier. Second, the decoder reconstructed the original inputs with the latent and class vectors. Third, we imposed a categorical distribution on the discrete class vectors represented in the one-hot form using the adversarial regularization process and incorporated the supervised classification process into the network to better guide the network training using the limited labeled data. The comparison experiments on the synthetic dataset and measured hyperspectral dataset were conducted to quantitatively and qualitatively certify the effect of the proposed method. The results show that the SSAAE outperformed six other methods for classifying alteration zones. Moreover, we further displayed the delineated results of the SSAAE on the cross-section, in which the alteration zones were sensible from a geological point of view and had good spatial consistency with the occurrence of Cu, which further demonstrates that the SSAAE had good applicability for the classification of alteration zones.

Modeling and detection of invasive trees using UAV image and machine learning in a subtropical forest in Brazil

Protected areas play an extremely important role in the conservation of global biodiversity. However, these areas are subject to the introduction of invasive alien species (IAS), which cause damage to native environments. The present study aimed to use images obtained by Unmanned Aerial Vehicles (UAVs) combined with machine learning (ML) algorithms to identify the IAS Hovenia dulcis in a Conservation Unit in southern Brazil. Field data were obtained in a sample area, where the floristic survey of the H. dulcis species was carried out. To obtain remote data, a UAV with a built-in RGB sensor was used. Subsequently, the images were processed for orthomosaic generation and the spatial distribution of the inventoried species, based on manual photointerpretation. Furthermore, in the supervised classification process, four classes of interest were defined: H. dulcis, similar species, shade, and other species. The process involved two approaches (pixel-based - PB and object-based image analysis - OBIA) and two ML algorithms were compared (Random Forest - RF and Support Vector Machine - SVM). Samples were separated into 90% for training and 10% for model validation. For performance analysis, overall accuracy (OA) and Kappa index metrics were calculated. The results show that the RF algorithm in the PB approach had the best performance in the classification of the IAS H. dulcis, presenting a kappa of 0.87 and OA of 91.5%, in the training data set and 90.91% of success in the model validation dataset. Our study demonstrated to be able to reach the results to respond to the raised hypotheses. Furthermore, the UAV-RGB data combined with ML are highly accurate to identify H. dulcis in relation to the other species that make up the forest stratum of the study area.

Spatial - Temporal variability of Land Use Land Cover at Mount Merapi, Indonesia

One of the most active volcanoes in the island of Java is Merapi mount which was experienced the last major eruption peak on October 26th, 2010. This volcanic eruption was effusive eruption type where magmatic gas pressure in the crater was not too strong and magma eruption was just flown out past the slopes of the Merapi mount area. However, magmatic gas pressure and magma volume still result in deformation changes that have a direct impact on residential areas throughout the Merapi mount area. Residential areas were obtained through supervised classification process from Landsat 7 and 8-satellite imagery in the 2009, 2011 and 2019 acquisition year. The reason of observation year selection was based on pre and post eruption concept to get pattern of Merapi’s mountain body change through deformation analysis. The work focuses on spatial-temporal variability of land use land cover analysis at Mount Merapi pre and post 2010 eruption event considered here. The technique is based on NDVI (Normalized Difference Vegetation Index), Maximum Likelihood Classification (MLC) and False Colour Composite methodology. Based on change in number of pixels it was analysed. Actually, some portion of land was covered with clouds and its shadows. From the results it was observed that, water body, barren and built up features were miss classified. So finally False Colour Composite (FCC) images are used to identify the misclassified classes.

Evaluating the Territorial Impact of Built-Up Area Expansion in the Surroundings of Bucharest (Romania) through a Multilevel Approach Based on Landsat Satellite Imagery

Assessing the relentless expansion of built-up areas is one of the most important tasks for achieving sustainable planning and supporting decision-making on the regional and local level. In this context, techniques based on remote sensing can play a crucial role in monitoring the fast rhythm of urban growth, allowing the regular appraisal of territorial dynamics. The main aim of the study is to evaluate, in a multi-scalar perspective, the built-up area expansion and the spatio–temporal changes in Ilfov County, which overlaps the surroundings of Bucharest, capital of Romania. Our research focuses on processing multi-date Landsat satellite imagery from three selected time references (2000, 2008, 2018) through the supervised classification process. Further on, the types of built-up area dynamics are explored using LDTtool, a landscape metrics instrument. The results reveal massive territorial restructuring in the 18 years, as the new built-up developments occupy a larger area than the settlements’ surface in 2000. The rhythm of the transformations also changed over time, denoting a significant acceleration after 2008, when 75% of the new development occurred. At the regional level, the spatial pattern has become more and more complex, in a patchwork of spatial arrangements characterized by the proliferation of low density areas interspersed with clusters of high density developments and undeveloped land. At the local level, a comparative assessment of the administrative territorial units’ pathway was conducted based on the annual growth of built-up areas, highlighting the most attractive places and the main territorial directions of development. In terms of the specific dynamics of built-up areas, the main change patterns are “F—NP increment by gain”, followed by “G—Aggregation by gain”, both comprising around 80% of the total number of cells. The first type was prevalent in the first period (2000–2008), while the second is identified only after 2008, when it became the most represented, followed in the hierarchy by the previously dominant category. The spatial pattern differentiations were further explored in three complementary case studies investigated in correlation with socioeconomic data, revealing a heterogeneous landscape.

UNESCO World Heritage properties in changing and dynamic environments: change detection methods using optical and radar satellite data

The article presents recent capabilities of active and passive earth observation sensors along with related processing image chains, for monitoring UNESCO World Heritage properties. Exceptional heritage sites and landscapes are found in dynamic environments, whereas both anthropogenic and natural changes are observed. The use of radar and optical satellite imageries can be used as a systematic observation tool for stakeholders, to map drastic or slowly driven landscape changes towards the better protection and management of these sites and their surrounding areas. The study presents the results from the analysis of the European Copernicus Sentinel-1 and Sentinel-2 satellite images over two broader areas in the Eastern Mediterranean basin that hold important UNESCO World Heritage properties. Initially, a recent strong earthquake of a 6.7 magnitude scale in the Aegean Sea is studied using radar Sentinel-1 images. These radar images were processed through the Hybrid Pluggable Processing Pipeline (HyP3) cloud platform for analyzing both significant changes of the VV (vertical transmit, vertical receive) and VH (vertical transmit, horizontal receive) backscattering signal as well as through an Interferometric Synthetic Aperture Radar (InSAR) analysis. Then, long-term changes in Cyprus during the last two decades are monitored by a Sentinel-2 image compared to the European Corine Land Use Land Cover data of 2000. These changes are mapped after a supervised classification process using the random forest (RF) classifier. The overall results demonstrate that the recent developments of the space sector in all its segments (resolution of the sensors, the capacity to storage in the cloud, processing advancements and open-access datasets and tools) can be beneficial for monitoring UNESCO World Heritage properties.

Geo-Object-Based Vegetation Mapping via Machine Learning Methods with an Intelligent Sample Collection Scheme: A Case Study of Taibai Mountain, China

Precise vegetation maps of mountainous areas are of great significance to grasp the situation of an ecological environment and forest resources. In this paper, while multi-source geospatial data can generally be quickly obtained at present, to realize effective vegetation mapping in mountainous areas when samples are difficult to collect due to their perilous terrain and inaccessible deep forest, we propose a novel and intelligent method of sample collection for machine-learning (ML)-based vegetation mapping. First, we employ geo-objects (i.e., polygons) from topographic partitioning and constrained segmentation as basic mapping units and formalize the problem as a supervised classification process using ML algorithms. Second, a previously available vegetation map with rough-scale label information is overlaid on the geo-object-level polygons, and candidate geo-object-based samples can be identified when all the grids’ labels of vegetation types within the geo-objects are the same. Third, various kinds of geo-object-level features are extracted according to high-spatial-resolution remote sensing (HSR-RS) images and multi-source geospatial data. Some unreliable geo-object-based samples are rejected in the candidate set by comparing their features and the rules based on local expert knowledge. Finally, based on these automatically collected samples, we train the model using a random forest (RF)-based algorithm and classify all the geo-objects with labels of vegetation types. A case experiment of Taibai Mountain in China shows that the methodology has the ability to achieve good vegetation mapping results with the rapid and convenient sample collection scheme. The map with a finer geographic distribution pattern of vegetation could clearly promote the vegetation resources investigation and monitoring of the study area; thus, the methodological framework is worth popularizing in the mapping areas such as mountainous regions where the field survey sampling is difficult to implement.

Analysis of the Impacts of Urbanisation on Land Cover Change in Gombe Local Government Area, Gombe State Nigeria

Like many other Nigerian capital cities of newly created states in the 1990s, Gombe the capital city of Gombe State in the North-eastern geopolitical zone has experienced tremendous change in the land cover which is primarily driven by urbanisation. However, little is known on the extent of this urban expansion, how it has affected other land cover types or the dynamics of demographic change as a function of urban growth in the area. This spatial information is highly needed for effective planning and development. This paper, therefore, attempts to answer these questions by firstly, processing LandSat5 images of 1998 and Landsat 8 images of 2016 both of November of Gombe Local Government Area where the state capital is situated. Image processing is then carried out using the semi-automated classification plugin in QGIS 2.18. A supervised classification scheme was used to classify the 1998 and 2016 Landsat Image scenes into four land cover classes (water bodies, built-up area, bare-surface and vegetation) using the spectral angle mapping algorithm. Secondly, the paper analysed population data of Gombe Local Government Area for the same period to understand the dynamics of population change in the area. Based on the findings, urban land cover type increased from 13.02 Km2 (25.14% of the total land cover) in November 1998 to 25.98 km2 (50.17% of the entire land cover) in November 2016. As a result of this change in land cover, all other land cover types decreased in areal coverage. A Kappa index of 0.82 and 0.81 suggest that that the error margins during the supervised classification process of the 1998 and 2016 Landsat images are relatively small. The implications of rapid changes in land cover and population change in the area over a short period of 18 years on planning and management are also discussed.

Inland Lakes Mapping for Monitoring Water Quality Using a Detail/Smoothing-Balanced Conditional Random Field Based on Landsat-8/Levels Data.

The sustainable development of water resources is always emphasized in China, and a set of perfect standards for the division of inland water environment quality have been established to monitor water quality. However, most of the 24 indicators that determine the water quality level in the standards are non-optically active parameters. The weak optical characteristics make it difficult to find significant correlations between the single parameters and the remote sensing imagery. In addition, traditional on-site testing methods have been unable to meet the increasingly extensive water-quality monitoring requirements. Based on the above questions, it’s meaningful that the supervised classification process of a detail-preserving smoothing classifier based on conditional random field (CRF) and Landsat-8 data was proposed in the two study areas around Wuhan and Huangshi in Hubei Province. The random forest classifier was selected to model the association potential of the CRF. The results (the first study area: OA = 89.50%, Kappa = 0.841; the second study area: OA = 90.35%, Kappa = 0.868) showed that the water-quality monitoring based on CRF model is feasible, and this approach can provide a reference for water-quality mapping of inland lakes. In the future, it may only require a small amount of on-site sampling to achieve the identification of the water quality levels of inland lakes across a large area of China.

The study of NDVI fluctuation in southern Iraq(Hor Ibn Najim) using remote sensing data

This study was carried out to determine the nature of the land cover in the area of Hor Ibn Najim and its surroundings. These changes had occurred changes in the environmental conditions between 1986 and 2016. Remote sensing(RS) and geographic information system (GIS) techniques were used to monitor changes in the region as well as changes in NDVI values. Satellite images of Landsat 5 and 7 with sensors TM and ETM +, were used respectively. The extraction of land covers in the region was executed by the application of the supervised classification processes. The area has been classified in to six main land categories (water bodies, low density of vegetation cover, high density of vegetation cover, wetlands, abandoned land and land suitable of agriculture).

The Impacts Analysis of Pre And Post Merapi Mount Eruption on Residential Areas Using Sentinel 1, ALOS Palsar and Landsat Satellite Images Combination in 2009-2015

Indonesia is a country located in the presence of three large tectonic plates like the Indo-Australian, Eurasian and Pacific plates. The path of plate movement approaching or steering away each other leads to high volcanic and earthquakes activity. As a proof of the active path of this plate movement can be seen from the series of active volcanoes lies from the island of Sumatra, Java, Nusa Tenggara into Papua. The impact of active volcano activity will be felt in the densely populated area where the island of Java became one of the islands with the largest residential area in Indonesia. This impact needs to be anticipated with an integrated high-tech technology and volcanic disaster mitigation system. One of the most active volcanoes in the island of Java is Merapi mount which is experienced the last major eruption peak on October 26th, 2010. This volcanic eruption is effusive eruption type where magmatic gas pressure in the crater is not too strong and magma eruption is just flowing out past the slopes of the Merapi mount area. However, magmatic gas pressure and magma volume still result in deformation changes that have a direct impact on residential areas throughout the Merapi mount area. Therefore, this research performs deformation calculation using DInSAR (Differential Interferometry Synthetic Aperture Radar) method on Sentinel 1 SAR and ALOS PALSAR 1 satellite images between 2005-2015 acquisition year. Then the deformation patterns obtained from the DInSAR method will be spatially correlated with residential areas throughout the Merapi mount area. Residential areas were obtained through supervised classification process from Landsat 5 and 8-satellite imagery in the 2005-2015 acquisition year. The reason of observation year selection is based on pre and post eruption concept to get pattern of Merapi’s mountain body change through deformation analysis. The result of the combination of deformation pattern and residential area changes was further analysed to obtain spatial correlation. The value of this spatial correlation will be an indicator of how’s far the deformation changes due to volcanic activity will have an effect on the residential areas in the Merapi mount area. The output of this research is a map of Merapi mount deformation (due to volcanic activity) impacts on the residential area. The results are expected to contribute to the volcanic disaster mitigation system and the development of residential area plans in Merapi mount area.

Low-illuminated SPOT-5 image improvement for density-based vegetation identification using three-layer colour manipulation approach

Poor illumination quality of a satellite image is one of the challenges encountered in vegetation analysis, especially with regard to pan-sharpened medium spatial resolution SPOT-5 imagery. Hence, the accuracy of vegetation identification will be affected. In this paper, a three-layer colour manipulation approach is proposed to overcome this issue of low illuminated SPOT-5 images in order to increase the performance of precise vegetation identification. The SPOT-5 image is pre-processed and three layers of image enhancement techniques are used to, specifically: identify vegetation, reduce shadow appearance, as well as contrast enhancement for colour uniformities in order to improve low illumination quality of images. These steps are then followed by a supervised classification process for density-based vegetation area discrimination. This research was tested using multispectral medium spatial resolution SPOT-5 imagery covering the Ramsar convention site of Tanjung Piai located at the southernmost tip of mainland Asia over the years 2008, 2011 and 2013. The results showed that the proposed approach performed better than existing techniques when dealing with low-illuminated medium resolution multispectral imagery specifically with regard to density-based vegetation identification. The results are supported with accuracy assessments and ground truth validation.

Satellite Image Classification Method Using the Dempster–Shafer Approach

In this paper the problem of the supervised classification of satellite images is considered. A new image classification method focused on application under conditions when the training sample is (in particular, considerably) contaminated is proposed. The method is based on using the Dempster–Shafer evidence theory and is applicable both for hyperspectral and multispectral satellite images. Problems of organizing the supervised classification process and content of its constitutive procedures are presented. The developed method has been implemented algorithmically and in software. Results obtained in the classification of hyperspectral images by the proposed method testify to its efficiency.

Satellite Images Classification in Rural Areas Based on Fractal Dimension

Fractal geometry is receiving increase attention as a quantitative and qualitative model for natural phenomena description, which can establish an active classification technique when applied on satellite images. In this paper, a satellite image is used which was taken by Quick Bird that contains different visible classes. After pre-processing, this image passes through two stages: segmentation and classification. The segmentation carried out by hybrid two methods used to produce effective results; the two methods are Quadtree method that operated inside Horizontal-Vertical method. The hybrid method is segmented the image into two rectangular blocks, either horizontally or vertically depending on spectral uniformity criterion; otherwise the block is segmented by the quadtree. Then, supervised classification is carried out by means the Fractal Dimension. For each block in the image, the Fractal Dimension was determined and used to classify the target part of image. The supervised classification process delivered five deferent classes were clearly appeared in the target part of image. The supervised classification produced about 97% classification score, which ensures that the adopted fractal feature was able to recognize different classes found in the image with high accuracy level.

Riverbank Filtration Site Suitability Selection Using Spatial Data Techniques: Case Study for Kota Lama Kiri, Kuala Kangsar

Site surface characterization is an important factor to identify a suitable area for riverbank filtration (RBF) location. However, selecting the suitable area for RBF location using conventional methods is costly and time consuming, usually restricted to a small area. In this research, a site suitability for RBF location methodology was proposed using spatial data techniques to determine the site suitability of the potential RBF location in Kota Lama Kiri, Kuala Kangsar study area. A high resolution GeoEye-1 satellite imagery acquired in 2012 was classified using the supervised classification process for land cover. The classified image was further analyze using overlaying, buffering and Boolean analysis, to identify the suitable site for RBF based on location, distance from the river and distant from built-up area. In addition, the geology and hydrological data were extracted from published maps, which were then converted and integrated into GIS spatial database. The results show that the classified GeoEye-1 image produces the overall accuracies of 83.50% % with kappa statistic value of 0.806. The site suitability map for the potential RBF locations in the study area were produced confirms the location of an existing RBF well developed by Lembaga Air Perak (LAP). The methodology can be readily used to provide information of suitability area for RBF location in which can be used by water supply management to locate the RBF well for extraction purposes.

Testing methods for using high-resolution satellite imagery to monitor polar bear abundance and distribution

High-resolution satellite imagery is a promising tool for providing coarse information about polar species abundance and distribution, but current applications are limited. With polar bears (Ursus maritimus), the technique has only proven effective on landscapes with little topographic relief that are devoid of snow and ice, and time-consuming manual review of imagery is required to identify bears. Here, we evaluated mechanisms to further develop methods for satellite imagery by examining data from Rowley Island, Canada. We attempted to automate and expedite detection via a supervised spectral classification and image differencing to expedite image review. We also assessed what proportion of a region should be sampled to obtain reliable estimates of density and abundance. Although the spectral signature of polar bears differed from nontarget objects, these differences were insufficient to yield useful results via a supervised classification process. Conversely, automated image differencing—or subtracting one image from another—correctly identified nearly 90% of polar bear locations. This technique, however, also yielded false positives, suggesting that manual review will still be required to confirm polar bear locations. On Rowley Island, bear distribution approximated a Poisson distribution across a range of plot sizes, and resampling suggests that sampling >50% of the site facilitates reliable estimation of density (CV <15%). Satellite imagery may be an effective monitoring tool in certain areas, but large-scale applications remain limited because of the challenges in automation and the limited environments in which the method can be effectively applied. Improvements in resolution may expand opportunities for its future uses. © 2015 The Wildlife Society.

Semantic localization in the PCL library

The semantic localization problem in robotics consists in determining the place where a robot is located by means of semantic categories. The problem is usually addressed as a supervised classification process, where input data correspond to robot perceptions while classes to semantic categories, like kitchen or corridor.In this paper we propose a framework, implemented in the PCL library, which provides a set of valuable tools to easily develop and evaluate semantic localization systems. The implementation includes the generation of 3D global descriptors following a Bag-of-Words approach. This allows the generation of fixed-dimensionality descriptors from any type of keypoint detector and feature extractor combinations. The framework has been designed, structured and implemented to be easily extended with different keypoint detectors, feature extractors as well as classification models.The proposed framework has also been used to evaluate the performance of a set of already implemented descriptors, when used as input for a specific semantic localization system. The obtained results are discussed paying special attention to the internal parameters of the BoW descriptor generation process. Moreover, we also review the combination of some keypoint detectors with different 3D descriptor generation techniques.

Satellite monitoring of land-use and land-cover changes in northern Togo protected areas

Remote-sensing data for protected areas in northern Togo, obtained in three different years (2007, 2000, and 1987), were used to assess and map changes in land cover and land use for this drought prone zone. The normalized difference vegetation index (NDVI) was applied to the images to map changes in vegetation. An unsupervised classification, followed by classes recoding, filtering, identifications, area computing and post-classification process were applied to the composite of the three years of NDVI images. Maximum likelihood classification was applied to the 2007 image (ETM+2007) using a supervised classification process. Seven vegetation classes were defined from training data sets. The seven classes included the following biomes: riparian forest, dry forest, flooded vegetation, wooded savanna, fallows, parkland, and water. For these classes, the overall accuracy and the overall kappa statistic for the classified map were 72.5% and 0.67, respectively. Data analyses indicated a great change in land resources; especially between 1987 and 2000 probably due to the impact of democratization process social, economic, and political disorder from 1990. Wide-scale loss of vegetation occurred during this period. However, areas of vegetation clearing and regrowth were more visible between 2000 and 2007. The main source of confusion in the contingency matrix was due to heterogeneity within certain classes. It could also be due to spectral homogeneity among the classes. This research provides a baseline for future ecological landscape research and for the next management program in the area.