Urban forests provide vital services to communities and are crucial for our mental, physical and emotional well-being. Recent research has shown that many variables at a neighbourhood-level are linked to variations in urban forest quantity, however, relationships at the property-level have not been considered. The purpose of this study was to examine the relationships at property-level in four socioeconomically varied neighbourhoods in the City of Mississauga (Ontario, Canada). Percent canopy cover and tree density was calculated using information from a survey, GIS datasets and remote sensing. Regression was used to determine which property-level characteristics are related to variations in the two tree cover variables. The results show that variables dealing with residents attitudes towards trees and space constraints are commonly linked to tree cover variations. The study found differences in relationships between the two tree measures at property-level and it provides greater insight into human-urban forest relationship at the micro-scale.

Cape Bounty, Melville Island is a partially vegetated High Arctic landscape with three main plant communities: polar semi-desert (47% of the study area), mesic tundra (31%) , and wet sedge meadows (7%). The objective of this research was to relate biophysical measurements of soil, vegetation, and CO2 exchange rates in each vegetation type to high resolution satellite data from IKONOS-2, extending plot level measurements to a landscape scale. Field data was collected through six weeks of the 2008 growing season. Two IKONOS images were acquired, one on July 4th and the other on August 2nd. Two products were generated from the satellite data: a land-cover classification and the Normalized Difference Vegetation Index (NDVI). The three vegetation types were found to have distinct soil and vegetation characteristics. Only the wet sedge meadows were a net sink for CO2; soil respiration tended to exceed photosynthesis in the sparsely vegetated mesic tundra and polar semi-desert. Scaling up the plot measurements by vegetation type area suggested that Cape Bounty was a small net carbon source (0.34 ± 0.47 g C m-2 day-1) in the summer of 2008. NDVI was strongly correlated with percent vegetation cover, vegetation volume, soil moisture, and moderately with soil nitrogen, biomass, and leaf area index (LAI). Photosynthesis and respiration of CO2 both positively correlated with NDVI, most strongly when averaged over the season. NDVI increased over time in every vegetation type, but this change was not reflected in any significant measured changes in vegetation or CO2 flux rates. A simple spatial model was developed to estimate Net Ecosystem Exchange (NEE) at every pixel on the satellite images based on NDVI, temperature and incoming solar radiation. It was found that the rate of photosynthesis per unit NDVI was higher early in the growing season. The model estimated a mean flux to the atmosphere of 0.21 g C m-2 day-1 at the time of image acquisition on July 4th, and -0.07 g C m-2 day-1 (a net C sink) on August 2nd. The greatest uncertainty in the relationship between NDVI and CO2 flux was associated with the polar semi-desert class. / Thesis (Master, Geography) -- Queen's University, 2011-12-28 23:27:34.824

Urban forests provide vital services to communities and are crucial for our mental, physical and emotional well-being. Recent research has shown that many variables at a neighbourhood-level are linked to variations in urban forest quantity, however, relationships at the property-level have not been considered. The purpose of this study was to examine the relationships at property-level in four socioeconomically varied neighbourhoods in the City of Mississauga (Ontario, Canada). Percent canopy cover and tree density was calculated using information from a survey, GIS datasets and remote sensing. Regression was used to determine which property-level characteristics are related to variations in the two tree cover variables. The results show that variables dealing with residents attitudes towards trees and space constraints are commonly linked to tree cover variations. The study found differences in relationships between the two tree measures at property-level and it provides greater insight into human-urban forest relationship at the micro-scale.

Desde a década de 70, técnicas de Sensoriamento Remoto têm sido usadas para estudar os recursos naturais. A partir da década de 90, o lançamento de satélites com sensores de alta resolução levou à implementação de novas abordagens no processamento digital de imagens. O mapeamento florestal é uma das ferramentas fundamentais para efetuar o diagnóstico atual da situação de remanescentes florestais, possibilitando, assim, a definição de estratégias que poderão conciliar a conservação da natureza e o desenvolvimento econômico de uma propriedade ou região. A presente pesquisa avaliou a possibilidade de se identificar e discriminar tipologias florestais presentes em um fragmento de Floresta Ombrófila Mista, visando desenvolver uma metodologia para o mapeamento dos remanescentes deste ecossistema de forma rápida, com baixo custo e precisão aceitável. Neste estudo de caso, os trabalhos foram realizados na Reserva Florestal EMBRAPA/EPAGRl, localizada no município de Caçador-SC. Levantamentos de campo, através do apontamento de áreas consideradas "alvo", forneceram um importante suporte para a definição das tipologias presentes na área de estudo. Em imagem Ikonos foram aplicadas técnicas de processamento digital de imagens, testando-se algoritmos de segmentação e classificação orientada a regiões como ferramentas para a descrição do estado atual da floresta. A avaliação do processamento digital foi efetuada comparando-se os resultados com um mapa-, referência ("verdade de campo"), elaborado no software ArcView, por meio de interpretação visual da imagem via tela do monitor. A definição das classes de mapeamento foi baseada na presença de espécies indicadoras de estágio sucessional com fisionomia arbustiva a arbórea que compõem o dossel da floresta. O mapeamento foi estruturado em dois níveis hierárquicos, sendo o primeiro referente aos estágios sucessionais e, o segundo, às tipologias propriamente ditas. Foram definidas e mapeadas 13 classes temáticas por meio de interpretação visual, sendo oito destas referentes a tipologias florestais. Métodos de análises qualitativas e quantitativas foram empregados para a definição dos melhores pares de limiares para o processo de segmentação. Para tanto, na análise quantitativa desenvolveu-se uma modificação do índice IAVAS, denominado IAVASmod' Este índice permitiu comparar os diferentes limiares de similaridade e área, eliminando, com isto, a subjetividade de uma avaliação qualitativa na definição das melhores combinações de pares limiares. Dentre os pares testados, o melhor foi o par de limiares 35 para similaridade e 1.200 para área. As regiões geradas por este par de limiares foram submetidas ao processo de classificação, empregando-se os algoritmos "Isoseg" e "Bhattacharyya", disponíveis no software SPRlNG. Na classificação digital, reduziu-se para 11 o número de classes devido à não discriminação de uma classe referente à tipologia florestal e o agrupamento de outras duas referentes ao uso do solo. A classificação digital supervisionada apresentou-se eficiente para discriminar a tipologia "Predominância de Araucária". Para as demais tipologias florestais o classificador Bhattacharyya não demonstrou uma performance adequada, fato que influenciou os baixos valores de acuracidade geral (51,73%) e o índice kappa (0,43).

Orientador: Erivaldo Antonio da Silva / Resumo: O grande avanço tecnológico desta década, na área de Sensoriamento Remoto, pode ser percebido quando são observadas as grandes mudanças nas características dos sistemas orbitais mais tradicionais, bem como da nova geração de sistemas sensores desenvolvidos com o intuito de auxiliar, cada vez mais, as tarefas de identificação de alvos na superfície terrestre, devido à grande melhoria na resolução espacial. Produtos orbitais de alta resolução, com grau de detalhamento em torno do metro, permitem um melhor aproveitamento das imagens em aplicações cartográficas. O mercado de mapeamento urbano atualmente é ainda quase inteiramente baseado em fotografias aéreas. Porém, o Sensoriamento Remoto orbital vem passando por uma grande evolução tecnológica desde o final de 1999, quando foi lançado pela empresa norte-americana Space Imaging o satélite IKONOS. Este satélite possui sensores capazes de gerar imagens com lmetro de resolução espacial no modo pancromático e 4 metros no modo multiespectral. Estas imagens permitem o mapeamento da cobertura e uso do solo de maneira detalhada e continuada, desde que sejam usados métodos e/ou técnicas apropriadas. Este trabalho teve como objetivo fazer um estudo do potencial de uso das imagens geradas pelo satélite IKONOS, produto Geo, no que diz respeito a escala máxima de utilização em aplicações cartográficas. O procedimento para verificar a exatidão cartográfica baseou-se na análise estatística das discrepâncias entre as coordenadas de pontos no terreno, obtidas através do GPS, e as coordenadas dos pontos homólogos extraídas da imagem IKONOS, através da análise da existência de tendências e da precisão. Como resultado final, chegou-se a conclusão que a imagem IKONOS/Geo utilizada é adequada a escala 1:50000 e menores. / Abstract: The huge technological advancement that occurred in this decade, in the field of Remote Sensing, can be well perceived when we observe the great changes that occurred in the characteristics of the more traditional orbital systems, as well as of those which belong to the new generation of sensor systems developed with the aim of helping, more and more, the tasks of identification of targets on the Earth surface, due to the improvement on the spatial resolution. Orbital products of high resolution with the possibility of showing details of about one meter in size allow a better employment of imagery in cartographic applications. The urban mapping market is nowadays almost totally based on aerial photography. However, the orbital Remote Sensing is getting through a immense technological evolution since the end of 1999, when the satellite IKONOS was launched by a north American company called Space Imaging. This satellite has sensors capable of generating images with 1 meter resolution in the panchromatic mode and 4 meter resolution in the multispectral mode. These imagery allow mapping the land cover and use in a detailed and continuous manner, providing the appropriate methods and/or techniques are used. This dissertation aimed at studying the potential use of such imagery obtained by IKONOS satellite, Geo Product, specially with respect to the maximum scale of employment for cartographic applications. The approach for the checking the cartographic accuracy was based upon the statistical analysis of discrepancies between the coordinates on the ground, obtained by the use of GPS, and the coordinates of homologue points extracted from the IKONOS imagery, through the analysis of existence of trend and also by the analysis of precision. As a final result, it has been found that the IKONOS/Geo imagery is useful for mapping at 1:50.000 and smaller scales. / Mestre

Résumé : L’ombre sur l’imagerie satellitaire à très haute résolution spatiale a suscité relativement peu de travaux de recherche en télédétection. La présente étude a pour objectif la détection des zones d’ombre sur l’imagerie IKONOS via l’exploitation des caractéristiques des couleurs de l’ombre. Ces caractéristiques découlent des couleurs photométriques invariantes issues des transformées couleurs. L’approche proposée a pour but d’évaluer dans un premier temps, le potentiel des transformées RGB normalisé, IHS (Intensity, Hue, Saturation), HSV (Hue, Saturation, Value), C[indice inférieur 1] C[indice inférieur 2] C[indice inférieur 3], L[indice inférieur 1] L[indice inférieur 2] L[indice inférieur 3] et M[indice inférieur 1] M[indice inférieur 2] M[indice inférieur 3] (Gevers and Smeulders, 1999) à discriminer les zones d’ombre. Cette évaluation s’est effectuée sur deux images simples dont la première est une image d’extérieur affectée par les effets atmosphériques, et la deuxième est une image d’intérieur, illuminée par une source de lumière artificielle. Les effets atmosphériques, notamment la dispersion, conduisent à la saturation maximale des zones d’ombre sur l’image d’extérieur. Cette conclusion a été vérifiée sur l’image IKONOS et exploitée pour développer une approche pour la détection de l’ombre sur ce genre de données. La transformée IHS a été retenue pour générer les couleurs photométriques H et S respectivement la teinte et la saturation. Deux dérivées I-S et H+l/I+1 ont été calculées. Ces dernières permettent le rehaussement des zones d’ombre sur l’image IKONOS. Trois méthodes de seuillage ont été appliquées sur les deux dérivées afin de particulariser les régions d’ombre. Les deux images finales issues du seuillage ont été fusionnées. La dernière étape de la recherche a consisté en une validation des résultats. Cette opération a démontré la robustesse de l’approche avec une précision globale moyenne de 80%. La confusion entre les zones d’ombre et les régions sombres est la principale faiblesse de l’approche proposée. Cependant, son amélioration est possible grâce à l’intégration d’autres types d’informations comme la texture et la prise en compte du voisinage. // Abstract : The shadow phenomena on the satellite imagery with very high spatial resolution has been the subject of a few research tasks in remote sensing. The aim of the present study is the detection of the shadow’s areas on IKONOS imagery with the use of the shadow’s colors characteristics. These characteristics were derived from the invariant photometric colors. The purpose of the suggested approach is to evaluate initially, the potential of normalized color RGB, IHS (Intensity, Hue, Saturation), HSV (Hue, Saturation, Value), C[subscript 1] C[subscript 2] C[subscript 3], L[subscript 1] L[subscript 2] L[subscript 3] et M[subscript 1] M[subscript 2] M[subscript 3] (Gevers and Smeulders, 1999) transform colors to discriminate shadow’s areas. This evaluation was carried out on two simple images whose first is an outside scene affected by the atmospheric effects, and the second is an interior one illuminated by artificial light source. The atmospheric effects, in particular dispersion, lead to the maximum saturation of shadow’s zones on the outside image. This conclusion was checked on IKONOS image and exploited to develop a method to detect the shadow on this kind of data. IHS transform was retained to generate the photometric colors hue (H) and saturation (S). Two derived images I-S and H+l/I+1 were calculated. They allow the raising of the shadow’s areas on IKONOS image. Three methods of thresholding were applied to the two derived images in order to differentiate the areas of interest. The two final images resulting from the thresholding were joined. The last stage of research consisted of a validation of the results. This operation showed the robustness of the approach with 80% of precision. Confusion between the shadow’s areas and the dark ones is the principal weakness of the suggested approach. However, this new technique can be improved by introducing other information like texture and nearest neighbor analysis.

Dans les régions de haute biodiversité, caractérisées par des paysages dynamiques, la cartographie détaillée de l'utilisation des sols et du couvert végétal est communément obtenue par la classification d'images satellite. Cependant, les cadres conceptuels d'estimation d'erreurs sur les cartes sont éprouvés pour les zones tempérées et hautement industrialisées.<br />Une nouvelle méthode est proposée pour l'évaluation de la fiabilité des cartes et une autre méthode pour l'estimation des erreurs de classification par ambiguïtés entre classes sur images satellites. La première méthode comprend un nouveau mode d'échantillonnage et une estimation par ensembles flous des incertitudes positionnelles et thématiques. Elle a été testée sur l'Inventaire Forestier Mexicain de l'an 2000. La deuxième méthode s'appuie sur la simulation d'images satellites avec le modèle de transfert radiatif DART et a été testée sur des images IKONOS de six types de forêts au Mexique, sur terrain plat et en forte pente

Polygon-based classification applied on the unitemporal SPOT4 XS, SPOT5 XS, IKONOS XS, QuickBird XS and QuickBird Pansharpaned (PS) images is described. The study site is an agricultural area located near Karacabey, Turkey covering an area of about 95 km2. The objective was to assess the effect of the spatial resolution on the polygon-based classification of agricultural crops. Both the post-polygon and pre-polygon classifications were carried out. In the post-polygon classification, first, the images were classified on per-pixel basis through a Maximum Likelihood classifier. Then, for each field, the model class was computed and the field was assigned the label of the model class. In the pre-polygon classification, first, the mean values were calculated for each field. Then, the per-pixel Maximum Likelihood Classification was carried out using the mean bands. The post-polygon classification of the SPOT4 XS and SPOT5 XS images produced an overall accuracy of 76,1% and 81,4%, respectively. The IKONOS XS image provided the highest overall accuracy of 88,6%. On the other hand, the QuickBird XS and QuickBird PS images provided the overall accuracies of 83,7% and 85,8%, respectively. For the pre-polygon classification, the overall accuracies of the SPOT4 XS and SPOT5 XS images were computed to be 65,2% and 69,8%, respectively. Similar to the post-polygon classification, the IKONOS image provided the highest overall accuracy of 81,8% while the SPOT5 XS image revealed slightly better results than the SPOT4 XS image. The overall accuracies of the QuickBird XS and PS images were found to be 78,6% and 82,1%, respectively.

No description available.

Soil landscape characterization into landform elements for precision agriculture has become an important issue. As soil properties and crop yields change over the landscape, delineating landform elements as a basis for site-specific application of crop inputs has become a reality. Two different methods of delineating landform elements from agricultural fields were tested and compared. The first method delineated landform elements from digital elevation maps with the use of the LandMapR(tm) software, the second method delineated classes from IKONOS high resolution panchromatic images using an unsupervised classification algorithm. The LandMapR(tm) model delineated landform elements from true elevation data collected in the field and was considered the reference dataset to which the image classification maps were compared to. The IKONOS imagery was processed using a combination of one filtering algorithm and one unsupervised classification method prior to being compared to the classified DEM. A total of 20 filtering algorithms and two unsupervised methods were used for each of the five study sites. The study sites consisted of four agricultural fields covered with crop stubble and one field in summer fallow. Image classification accuracy assessment was reported as overall, producer’s and user’s accuracy as well as Kappa statistic. Results showed that filtering algorithms and classification methods had no effects on image classification accuracies. Highest classification accuracy of image map to landform element map comparison achieved for all study sites was 17.9 %. Classification accuracy was affected by the heterogeneity of the ground surface cover found in each field. However, the classification accuracy of the fallow field was not superior to the stubble fields. / May 2006