Abstract

This paper illustrates how the capabilities of GIS and satellite imagery can be harnessed to explore and better understand the urban form of several large African cities (Addis Ababa, Nairobi, Kigali, Dar es Salaam, and Dakar). To allow for comparability across very diverse cities, this work looks at the above mentioned cities through the lens of several spatial indicators and relies heavily on data derived from satellite imagery. First, it focuses on understanding the distribution of population across the city, and more specifically how the variations in population density could be linked to transportation. Second, it takes a closer look at the land cover in each city using a semi-automated texture based land cover classification that identifies neighborhoods that appear more regular or irregularly planned. Lastly, for the higher resolution images, this work studies the changes in the land cover classes as one moves from the city core to the periphery. This work also explored the classification of slightly coarser resolution imagery which allowed analysis of a broader number of cities, sixteen, provided the lower cost. Document type: Book

Abstract

This chapter has sought to explain the context, policy applications, and major design choices in the process of developing an operational urban simulation model, with specific reference to UrbanSim as a case study. It has been argued that careful design at each stage of the process is needed to make the model sensitive to the policies of principal concern, to make the data and computational requirements manageable, to make the model usable by staff and other users with appropriate levels of training, and to fit into the operational practices of the relevant organizations. To be useful (relevant) in the policy process, model design should carefully integrate the elements discussed in the chapter into a design that fits well into a specific institutional and political context, and evolve to adapt to changing conditions. This introduction to the design process sets the stage for more in-depth discussion of specification and operational issues in model use. The UrbanSim system is being further developed to adapt to varying data availability, different factors influencing agent choices in locations ranging from newer and rapidly growing US metropolitan areas in other parts of the world. Considerable effort is now being devoted to developing environmental components of the system such as land cover change, and to developing a robust interface and tools for visualization and evaluation of policy scenarios. Document type: Part of book or chapter of book

Abstract

O transporte ferroviário de passageiros ganha importância estratégica na melhoria do trânsito e qualidade de vida dos grandes centros urbanos, mas no Brasil enfrenta sérios problemas operacionais, como obsolescência da infra-estrutura e invasão da faixa de domínio, comprometendo sua segurança e operacionalidade. O objetivo central deste trabalho é o mapeamento da cobertura do solo ao longo da faixa de domínio de uma linha de trem metropolitano. A área de estudo concentra-se nas adjacências da Linha F de trem metropolitano de São Paulo, operada pela CPTM Companhia Paulista de Trens Metropolitanos, com 38,8 km de extensão, cruzando as zonas central e leste da capital paulista, estendendo-se até outros municípios da Grande São Paulo. A metodologia adotada compreende o uso de imagem de satélite de alta resolução espacial IKONOS II, e técnica de classificação baseada em regiões. Os classificadores por regiões utilizam, além de informação espectral de cada pixel, a informação espacial que envolve a relação entre os pixels e seus vizinhos. Através da interpretação visual, foram identificadas doze classes de coberturas do solo, cujas feições nortearam os parâmetros da segmentação. A imagem segmentada foi então classificada usando o algoritmo de Bhattacharyya e pós-classificada utilizando regras e critérios definidos para cada classe, de acordo com seus atributos e contextualização. O desempenho da classificação final foi avaliado com base no coeficiente Kappa tendo atingido nível aceitável de correspondência. O mapa de cobertura do solo gerado a partir da classificação, demonstra que o entorno da linha é composto, predominantemente, por materiais impermeáveis como asfalto e concreto mas possui presença significativa de elementos naturais como vegetação e água. Desta forma, comprova-se a aplicabilidade de imagem de satélite de alta resolução e técnicas de processamento de imagem como suportes ao planejamento e gerenciamento do entorno da ferrovia. The rail transport of passengers has strategic importance in the traffic and quality of life improvement in the downtowns, but in Brazil it faces serious operational problems such as obsolete infrastructure and invasion of the right of way, compromising its safety and operability. The principal aim of this paper is the mapping of land cover area of metropolitan railway right of way. The study focuses on the outskirts of the Line F metropolitan train in Sao Paulo, operated by CPTM - Companhia Paulista de Trens Metropolitanos, with 38,8 km long, crossing the central and eastern areas of São Paulo, extending to other cities in the Metropolitan Region of Sao Paulo. The methodology includes the use of the high spatial resolution satellite image - IKONOS II, and regions-based classification. The regions-based classifiers use, besides spectral information of each pixel, the spatial information that involves the relationship between the pixels and its neighbours. Through visual interpretation, were identified twelve classes of land covers, which features guided the segmentation parameters. The segmented image was classified using Bhattacharyya distance and post-classified using rules and criteria established for each class, according to their attributes and contextualization. The performance of the final classification was evaluated based on Kappa coefficient, resulting acceptable level of correspondence. The mapping of land cover generated from the classification, shows that the outskirts of the metropolitan railway`s right way is predominantly composed by impermeable materials - such as asphalt and concrete - but has significant presence of natural elements - such as vegetation and water. Thus, is evident the applicability of high spatial resolution satellite image added to classification methods for remotely sensed data for planning and management of the railway right of way.