In this dissertation, methods are designed and validated for the utilization of hyperspectral hypertemporal remotely sensed data in target detection applications. Two new classes of methods are designed to optimize the selection of target detection features from spectro-temporal space data. The first method is based on the consideration that all the elements of the spectro-temporal map are independent of each other. The second method is based on the consideration that the elements of the spectro-temporal map have some vicinal dependency among them. Methods designed for these two approaches include various stepwise selection methods, windowing approaches, and clustering techniques. These techniques are compared to more traditional feature extraction methods such as Normalized Difference Vegetation Index (NDVI), spectral analysis, and Principal Component Analysis (PCA). The efficacies of the new methods are demonstrated within an aquatic invasive species detection application, namely discriminating waterhyacinth from other aquatic vegetation such as American lotus. These two aquatic plant species are chosen for testing the proposed methods as they have very similar physical characteristics and they represent a practical life target detection problem. It is observed from the overall classification accuracy estimates that the proposed feature extraction methods show a marked improvement over conventional methods. Along with improving the accuracy estimates, these methods demonstrate a capability to drastically reduce the dimensionality while retaining the desired hyperspectral hypertemporal features. Furthermore, the feature set extracted using the newly developed methods provide information about the optimum subset of the hyperspectral hypertemporal data for a specific target detection application, which makes these methods serve as tools to strategize more intelligent data collection plans.

First, the field survey campaigns have been carried out from March 22, 2007 to December 27, 2007 around 5-15 days in the interval in the study area of Nansha Island. The field work includes the survey of spatial distribution of various land use and crop types and the ground measurements of the crop biophysical parameters (such as the plant height, leave area index, fresh biomass, and plant water content) and the soil parameters (such as the soil water content and surface roughness parameters) of rice field and sugarcane field. And at the same time, the ENVISAT ASAR data were acquired from March 22, 2007 to December 27, 2007 in the interval of 35 days. During the acquisition dates of the ENVISAT ASAR data, the field surveys were also conducted. / Fourth, the sufficient ground measurements and simultaneous C-band HH- and VV-polarized SAR data of sugarcane crop have enriched the knowledge of understanding the temporal radar scatter mechanisms in sugarcane canopies. The C-band VV-polarized radar backscatters are larger than those of HH-polarization during the sugarcane growth cycle, and the difference is around 0.5 dB to 2 dB. The theoretical model MIMICS was adapted in modeling the scattering terms in sugarcane fields to interpret the temporal behavior of radar backscatters. For more robotic operation, the empirical regression models were used in estimation of the sugarcane LAI and fresh biomass, and mapping the sugarcane growth situation. The accuracies of the sugarcane LAI map and Biomass map are 0.74 and 0.70, respectively. / In conclusion, the C-band ENVISAT ASAR data can be efficiently used in the Pearl River Delta to monitor the crop growth, including the crop spatial distribution, crop acreages, and crop growth situation evaluation. The efficient crop growth monitoring program can not only help instruct the flexible farming actions, but also estimate the crop yield production for the decision-making government. (Abstract shortened by UMI.) / Second, field surveys were combined with the ENVISAT ASAR data to map the agricultural area. The analysis of the temporal radar backscatter characteristics of various land cover categories demonstrated that the time series of C-band SAR data is efficient in separating the eight land cover categories (rice paddy, sugarcane, banana, lotus ponds, mangrove wetlands, fish ponds, seawater, and buildings) in the PRD. The decision tree classifier is also approved to work efficiently on satellite SAR images with an overall accuracy of 77% and the Kappa coefficient of 0.74. The acreages of the land cover categories were also derived from the classification result with accuracies from 70% to 90%. / The Pearl River Delta is a typical developing region. It lies in the cloud-prone and rainy area of south China with multi-species of crops cultured in the agriculture areas. With a goal of developing an efficient, timely and accurate crop growth monitoring program in this area, field measurement, satellite SAR remote sensing technique, quantitative analysis of the crop biophysical parameters, and radar backscatter modeling methods have been integrated to study the multi-temporal and multi-polarized SAR data in estimating plant parameters (LAI, fresh biomass) of rice and sugarcane crop, and mapping the agricultural land cover categories of the study area in the PRD. / Third, in the study of rice growth monitoring, the trends of the relationships between C-band radar backscattering coefficients and rice parameters (plant height, LAI, fresh biomass, et al.) are proved to be constant with the reports in previous literatures. It was demonstrated that the differences between HH- and VV-polarized backscatter are not so evident (around 0.5 dB) in rice paddy canopies during the crop growth cycle. Moreover, by inducting a semi-empirical soil surface scattering component, a modified Water Cloud Model was developed to simulate the radar backscatter in rice crop canopies in different ground background situations (water surface, and soil surface) and to estimate the rice LAI and above ground fresh Biomass with reasonable accuracy. The rice growth conditions were displayed by LAI map and Biomass map generated from the model estimation, and the accuracies of the LAI and Biomass level classification are 0.77 and 0.71. / Wang, Dan. / Advisers: Hui Lin; Jin-Song Chen. / Source: Dissertation Abstracts International, Volume: 72-11, Section: B, page: . / Thesis (Ph.D.)--Chinese University of Hong Kong, 2009. / Includes bibliographical references (leaves 132-138). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [201-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract also in Chinese.

本研究以不同時期之航遙測影像偵測宜蘭海岸濱線變遷，影像來源包含1947年之舊航照影像、1971年的美國Corona衛星影像、1985年的像片基本圖、2003年的SPOT-5衛星影像及2009年以Z/I DMC(Digital Mapping Camera)航空數位相機所拍攝之高解像力航照影像。 由於影像獲取的時間與感測器皆有所差異，故本研究透過不同的方式處理資料，將影像地理對位，並利用地理資訊系統(Geographic Information Systems, GIS)軟體數化濱線及沙灘(丘)，且以套疊分析觀察不同時期濱線與沙灘變遷之情形，最後收集宜蘭地區的自然或人文資料如潮汐、降雨量與輸沙量等，分析宜蘭海岸變遷的原因。而在濱線萃取方面，由於以人工數化方式太耗時間與人力，故嘗試以半自動化方式如影像分類或影像分割萃取濱線，並與人工數化結果比較。研究結果顯示，利用多時期之遙測影像，並結合GIS之空間分析功能，確可有效掌握濱線與沙灘(丘)的歷史變化概況。 / This study used multi-temporal remote sensing images to detect shoreline changes along the Yilan coast. Various types of remote sensing images were used in this study, including old aerial images taken in 1947, Corona satellite images acquired in 1971, photo base map produced in 1985, SPOT-5 satellite images obtained in 2003, and high-resolution aerial images taken in 2009 by using Z/I DMC (Digital Mapping Camera). Because these images were taken in different time using different sensors, different procedures were applied to process the data and georeference the images to a common coordinate system. GIS (Geographic information Systems) software was used to digitize shoreline and the beach area, and overlay analysis was applied to find the shoreline changes in different time periods. Then various ancillary data such as tides, precipitation, and sediment load was collected to analyze the causes of coastal changes in Yilan. For shoreline extraction, manual digitization required a lot of time and manpower. Therefore, semi-automatic method such as image classification and image segmentation was applied to extract shoreline. The results show that, by using multi-temporal remote sensing images and spatial analysis functionalities of GIS, the historical changes of shoreline and beach area can be detected effectively.

MENVSC (Geography) / Department of Geography and Geo - Information Sciences / Detailed geomorphological mapping is important for monitoring environmental phenomena, it is therefore crucial that the methods employed for mapping are accurate. The basis of remote sensing for geomorphological work is moving from the consideration of whether satellite data are accurate for landform mapping to how surfaces of interest can be defined from remote sensing data, since earlier approaches of mapping are deemed costly and tedious. The aim of this study is to assess the suitability of ASTER and SRTM DEMs, and satellite imagery in detailed geomorphological mapping. Field survey and aerial photo interpretation were used to prepare a reference geomorphological map for comparisons. A similar approach of demarcating landform boundaries from aerial photographs was implemented to segment the DEMs into landform classes. The software packages that were used for processing the satellite data to create detailed geomorphological maps are QGIS with GRASS and SAGA plugins, and ENVI. The resultant geomorphological units’ maps from the DEMs when compared with the reference geomorphological map, show that the automated classification technique has advantages in terms of its efficiency and reproducibility. Nevertheless, distinct limitations of the technique are apparent and the technique is not suitable for detailed geomorphological mapping in the proposed study area. / NRF

Given satellite images with accompanying pixel classifications and elevation data, we propose different solutions to object detection. The first method uses hierarchical clustering for segmentation and then employs different methods of classification. One of these classification methods used domain knowledge to classify objects while the other used Support Vector Machines. Additionally, a combination of three Support Vector Machines were used in a hierarchical structure which out-performed the regular Support Vector Machine method in most of the evaluation metrics. The second approach is more conventional with different types of Convolutional Neural Networks. A segmentation network was used as well as a few detection networks and different fusions between these. The Convolutional Neural Network approach proved to be the better of the two in terms of precision and recall but the clustering approach was not far behind. This work was done using a relatively small amount of data which potentially could have impacted the results of the Machine Learning models in a negative way.

Documentation and formalization of informal settlements ("insitu" i.e. while people continue to live in the settlement) needs appropriate mapping and registration system of real property that can finally lead into integrating an informal city to the formal city. For many years extraction of geospatial data for informal settlement upgrading have been through the use of conventional mapping, which included manual plotting from aerial photographs and the use of classical surveying methods that has proved to be slow because of manual operation, very expensive, and requires well-trained personnel. The use of high-resolution satellite image like QuickBird and GIS tools has recently been gaining popularity to various aspects of urban mapping and planning, thereby opening-up new opportunities for efficient management of rapidly changing environment of informal settlements. This study was based on Manzese informal area in the city of Dar es salaam, Tanzania for which the Ministry of Lands and Human Settlement Development is committed at developing strategic information and decision making tools for upgrading informal areas using digital database, Orthophotos and Quickbird satellite image. A simple prototype approach developed in this study, that is, 'automatic detection and extraction of informal buildings and other urban features', is envisaged to simplify and speedup the process of land cover mapping that can be used by various governmental and private segments in our society. The proposed method, first tests the utility of high resolution QuickBird satellite image to classify the detailed 11 classes of informal buildings and other urban features using different image classification methods like the Box, maximum likelihood and minimum distance classifier, followed by segmentation and finally editing of feature outlines. The overall mapping accuracy achieved for detailed classification of urban land cover was 83%. The output demonstrates the potential application of the proposed approach for urban feature extraction and updating. The study constrains and recommendations for future work are also discussed. / Thesis (M.Env.Dev.)-University of KwaZulu-Natal, Pietermaritzburg, 2005.

[Truncated abstract] The assessment of rehabilitation success is time consuming and costly for bauxite miners because large areas of land (~550 ha per year) are involved. In some cases, rehabilitation results in patches of bare or sparsely vegetated soil. This study uses remote sensing imagery to evaluate the growth of vegetation in rehabilitated bauxite mines in the Darling Range, W.A. This work has five aims, which are to (1) compare vegetation biomass within rehabilitated areas and nearby native forest; (2) analyse temporal changes in vegetation growth within the selected rehabilitated areas, in particular rehabilitated areas with patches of bare soil; (3) compare vegetation growth pre- and post- mining; (4) identify the best type of remotely sensed data for this particular study area, and (5) develop an index, which can classify the degree of vegetation patchiness within rehabilitated mine sites. This information will enable rehabilitation workers to identify patches in rehabilitated areas that may require further remediation. The study used RADARSAT, nine years of Normalised Difference Vegetation Index (NDVI) maps (extracted from LANDSAT TM multivariate imagery and Quickbird imagery) and aerial photographs to evaluate forty-seven ~1 ha study sites. Image and map analyses were conducted mainly using ESRI’s software ArcGIS 8.3 and ER Mapper 6.4. Ground truthing was carried out to confirm and recognise the causes of bare patches within the rehabilitated mine sites ... The results indicate that differences in rehabilitation management do not affect this index but the extent of bare patches does. Due to the sensitivity of radar imagery to surface roughness, rehabilitated areas cannot be distinguished from the native forest using radar images. A building (crusher) appears to be the same as mature vegetation. Knowledge of the features in an area is therefore crucial when utilising RADARSAT. The beam elevation angle and profile of the RADARSAT image used, made superimposition of radar and optical imageries impossible. Speckle noise in RADARSAT images made it impossible to detect relatively small bare patches. In addition, the many cloud free days in Western Australia make optical imaging possible so that the ability of radar imagery to penetrate cloud is redundant for this type of study.

Abstract<p>Since 1999, Very High spatial Resolution satellite data (Ikonos-2, QuickBird and OrbView-3) represent the surface of the Earth with more detail. However, information extraction by multispectral pixel-based classification proves to have become more complex owing to the internal variability increase in the land-cover units and to the weakness of spectral resolution. <p>Therefore, one possibility is to consider the internal spectral variability of land-cover classes as a valuable source of spatial information that can be used as an additional clue in characterizing and identifying land cover. Moreover, the spatial resolution gap that existed between satellite images and aerial photographs has strongly decreased, and the features used in visual interpretation transposed to digital analysis (texture, morphology and context) can be used as additional information on top of spectral features for the land cover classification.<p>The difficulty of this approach is often to transpose the visual features to digital analysis.<p>To overcome this problem region-based classification could be used. Segmentation, before classification, produces regions that are more homogeneous in themselves than with nearby regions and represent discrete objects or areas in the image. Each region becomes then a unit analysis, which makes it possible to avoid much of the structural clutter and allows to measure and use a number of features on top of spectral features. These features can be the surface, the perimeter, the compactness, the degree and kind of texture. Segmentation is one of the only methods which ensures to measure the morphological features (surface, perimeter.) and the textural features on non-arbitrary neighbourhood. In the pixel-based methods, texture is calculated with mobile windows that smooth the boundaries between discrete land cover regions and create between-class texture. This between-class texture could cause an edge-effect in the classification.<p><p>In this context, our research focuses on the potential of land cover region-based classification of VHR satellite data through the study of the object extraction capacity of segmentation processes, and through the study of the relevance of region features for classifying the land-cover classes in different kinds of Belgian landscapes; always keeping in mind the parallel with the visual interpretation which remains the reference.<p><p>Firstly, the results of the assessment of four segmentation algorithms belonging to the two main segmentation categories (contour- and region-based segmentation methods) show that the contour detection methods are sensitive to local variability, which is precisely the problem that we want to overcome. Then, a pre-processing like a filter may be used, at the risk of losing a part of the information. The “region-growing” segmentation that uses the local variability in the segmentation process appears to be the best compromise for the segmentation of different kinds of landscape.<p>Secondly, the features calculated thanks to segmentation seem to be relevant to identify some land-cover classes in urban/sub-urban and rural areas. These relevant features are of the same type as the features selected visually, which shows that the region-based classification gets close to the visual interpretation. <p>The research shows the real usefulness of region-based classification in order to classify the land cover with VHR satellite data. Even in some cases where the features calculated thanks to the segmentation prove to be useless, the region-based classification has other advantages. Working with regions instead of pixels allows to avoid the salt-and-pepper effect and makes the GIS integration easier.<p>The research also highlights some problems that are independent from the region-based classification and are recursive in VHR satellite data, like shadows and the spatial resolution weakness for identifying some land-cover classes.<p><p>Résumé<p>Depuis 1999, les données satellitaires à très haute résolution spatiale (IKONOS-2, QuickBird and OrbView-3) représentent la surface de la terre avec plus de détail. Cependant, l’extraction d’information par une classification multispectrale par pixel devient plus complexe en raison de l’augmentation de la variabilité spectrale dans les unités d’occupation du sol et du manque de résolution spectrale de ces données. Cependant, une possibilité est de considérer cette variabilité spectrale comme une information spatiale utile pouvant être utilisée comme une information complémentaire dans la caractérisation de l’occupation du sol. De plus, de part la diminution de la différence de résolution spatiale qui existait entre les photographies aériennes et les images satellitaires, les caractéristiques (attributs) utilisées en interprétation visuelle transposées à l’analyse digitale (texture, morphologie and contexte) peuvent être utilisées comme information complémentaire en plus de l’information spectrale pour la classification de l’occupation du sol.<p><p>La difficulté de cette approche est la transposition des caractéristiques visuelles à l’analyse digitale. Pour résoudre ce problème la classification par région pourrait être utilisée. La segmentation, avant la classification, produit des régions qui sont plus homogène en elles-mêmes qu’avec les régions voisines et qui représentent des objets ou des aires dans l’image. Chaque région devient alors une unité d’analyse qui permet l’élimination de l’effet « poivre et sel » et permet de mesurer et d’utiliser de nombreuses caractéristiques en plus des caractéristiques spectrales. Ces caractéristiques peuvent être la surface, le périmètre, la compacité, la texture. La segmentation est une des seules méthodes qui permet le calcul des caractéristiques morphologiques (surface, périmètre, …) et des caractéristiques texturales sur un voisinage non-arbitraire. Avec les méthodes de classification par pixel, la texture est calculée avec des fenêtres mobiles qui lissent les limites entre les régions d’occupation du sol et créent une texture interclasse. Cette texture interclasse peut alors causer un effet de bord dans le résultat de la classification.<p><p>Dans ce contexte, la recherche s’est focalisée sur l’étude du potentiel de la classification par région de l’occupation du sol avec des images satellitaires à très haute résolution spatiale. Ce potentiel a été étudié par l’intermédiaire de l’étude des capacités d’extraction d’objet de la segmentation et par l’intermédiaire de l’étude de la pertinence des caractéristiques des régions pour la classification de l’occupation du sol dans différents paysages belges tant urbains que ruraux. / Doctorat en sciences agronomiques et ingénierie biologique / info:eu-repo/semantics/nonPublished

Ce travail s'inscrit dans un processus de réflexion qui vise à améliorer la compréhension de l'étalement urbain depuis la seconde guerre mondiale en périphérie bruxelloise, en particulier au travers de la cartographie et de l'analyse de l'utilisation du sol. Depuis cette date, l'urbanisation au sein des espaces ruraux à proximité des villes, processus qualifié de périurbanisation, s'est nettement renforcée. L'examen de la littérature dans ce domaine révèle que peu d'approches se basent sur l'emprise physique des modes d'utilisation du sol. C'est ce constat qui nous a amenée à envisager l'étude de la périurbanisation bruxelloise au travers des formes qu'elle a engendrées en se concentrant sur la notion d'utilisation du sol. Nous l'avons abordée au travers de trois aspects.<p><p>Tout d'abord, nous avons exploré l'apport des nouvelles techniques d'interprétation numérique à l'élaboration de cartes d'utilisation du sol à moyenne échelle (1:100 000) à partir d'images satellitaires. L'analyse des différentes étapes du processus de classification a montré que, lors de l'utilisation de classifications supervisées, la localisation et l'échantillonnage aléatoire des sites d'entraînement ainsi que la combinaison des caractéristiques des paramètres de classification (informations spectrales, texturales et contextuelles) améliorent considérablement l'exactitude du résultat. On constate également, dans ce dernier cas, l'importance de travailler avec une forme de voisinage isotrope et de pouvoir en faire varier la taille en fonction des classes considérées. L'utilisation de classificateurs multiples a permis de tendre vers une plus grande généralisation et de supprimer une série d'artefacts. Enfin, les essais ont montré que la classification par segmentation se rapproche fortement de la généralisation de l'interprétation visuelle tout en diminuant sensiblement le nombre d'objets à classer. <p><p>Dans un second temps, une analyse approfondie de la structure et de l’évolution de l’utilisation du sol à Bruxelles et dans sa périphérie au cours du dernier demi-siècle a été réalisée à partir d'une base de données diachronique à grande échelle (1:25 000). Elle confirme la vision d'une périphérie bruxelloise peu dense où le poids de la classe de tissu urbain résidentiel clairsemé discontinu s'intensifie au cours du temps. Les espaces privilégiés de l'urbanisation ont été identifiés de manière systématique à l'aide de canevas d'analyse radio-concentriques et directionnels. Le croisement entre les données d'utilisation du sol et d'autres indicateurs spatialisables comme les plans de secteur a montré les marges d'évolution potentielles, tant au moment de la mise en place de ces plans que plus récemment. D'autres combinaisons avec des statistiques socio-économiques ou démographiques ont fourni de nouveaux indicateurs permettant d'explorer les densités de l'occupation de l'espace périurbain et de s'interroger sur les discordances entre réalité physique et enregistrement statistique. Enfin, l'arrangement spatial des différentes classes d'utilisation du sol a été exploré. L'ensemble de ces analyses ont fait l'objet d'une synthèse thématique ainsi que régionale, au sein de compartiments paysagers.<p><p>La dernière approche de la problématique de l'utilisation du sol en milieu urbain s'est faite au travers d'une démarche basée sur la modélisation spatiale. La calibration des relations de voisinage à l'aide des règles d'autocorrélation spatiale a démontré que les affinités entre classes décrivent bien la structuration de l'agglomération bruxelloise et traduit le renforcement des structures héritées. L'analyse des résidus de la modélisation a montré le rôle contraignant de l'introduction des plans de secteur pour la classe du tissu urbain résidentiel clairsemé discontinu, laissant beaucoup plus de place qu'attendu dans les parties sud de la zone d'étude. La modélisation dynamique a aussi clairement mis en évidence le changement de logique de localisation de l'industrie et des services, qui rompent avec leurs localisations traditionnelles au cours de la période étudiée.<p> <p>En conclusion, la thèse a permis de confirmer les atouts d'une approche sous l'angle de l'utilisation du sol pour appréhender le phénomène de périurbanisation. Elle montre aussi le rôle unificateur de cette approche, qui peut s'insérer aisément dans les études thématiques ou susciter des questionnements nouveaux du fait des avancées apportées par le caractère quantitatif des exploitations régionales. Enfin, des outils communs et des enrichissements mutuels, acquis ou potentiels, sont identifiés entre les différents champs de la discipline (télédétection, géographie urbaine, modélisation spatiale) mobilisés pour cette exploration de la production, de l'analyse et de la modélisation des données d'utilisation du sol.<p> / Doctorat en sciences, Spécialisation géographie / info:eu-repo/semantics/nonPublished

Indiana University-Purdue University Indianapolis (IUPUI) / Urban forestry is a very important element of urban structures that can improve the environment and life quality within the urban areas. Having an accurate classification of urban forests and grass areas would help improve focused urban tree planting and urban heat wave mitigation efforts. This research project will compare the use of high – resolution aerial imagery and LiDAR data when used to classify canopy and grass areas. The high – resolution image, with 1 – meter resolution, was captured by The National Agriculture Imagery Program (NAIP) on 6/6/2012. Its coordinate system is the North American Datum of 1983 (NAD83). The LiDAR data, with 1.0 – meter average post spacing, was captured by Indiana Statewide Imagery and LiDAR Program from 03/13/2011 to 04/30/2012.The study area is called the Near East Side Community Organization (NESCO) neighborhood. It is located on the east side of downtown Indianapolis, Indiana. Its boundaries are: 65 interstate, East Massachusetts Avenue, East 21st Street, North Emerson Avenue, and the rail road tracks on the south of the East Washington Street. This research will also perform the accuracy assessment based on the results of classifications using high – resolution aerial imagery and LiDAR data in order to determine and explain which method is more accurate to classify urban canopy and grass areas.