Uso de imagens Radarsat-1 para fins de mapeamento de uso da terra e cobertura do solo em ambiente tropical úmido urbano: o caso de Manaus, Estado do Amazonas. / Use of radarsat-1 data in the identification of targets related to land use and land cover in urban humid tropical region: the case of Manaus, state of the Amazon, Brazil

AUTOR(ES)
DATA DE PUBLICAÇÃO

2002

RESUMO

The use of remote sensing data acquired through Synthetic Aperture Radar has been a recent common trend for land use and land cover mapping purposes. The information derived from this kind of data has been applied to distinct related application fields, such as: urban pattern analysis and mapping, settlements identification, urban change detection and even more specific tasks like population estimates and interpretation of socio-economic characteristics of urban areas. Within this context, the purpose of this investigation is general: to make a preliminary evaluation of the information derived from distinct modes of RADARSAT-1 images (Standard and Fine) acquired over the Manaus city during different passes (ascending and descending) for land use and land cover mapping. In addition, aspects related to the influence of sensor (incidence angle, look-azimuth and spatial resolution) and target (surficial roughness, moisture, type of terrain, etc) were also evaluated. The chosen test-site is the Manaus city, capital of Amazonas State, located in the center of the Brazilian Amazon Region. Five distinct images were acquired over the Manaus area under the GlobeSAR-2 Program encompassing descending Standard (S2, S5, S7) and a pair of opposite viewing Fine (F2, F2Near) images. The methodological approach can be divided into three main phases: 1) pre-processing, 2) visual analysis and 3) analysis of the results and integration. The first phase was related to the radiometric (speckle attenuation) and geometric correction (ortorretification) of the SAR dataset. The second phase has involved the visual analysis (photointerpretation) with the detection of the main targets and in the images related to urban land use (industrial sites, commercial sectors, residential areas, agricultural fields, infrastructure, communications, etc) and to land cover (drainage, vegetation, etc). The visual analysis was based on the classical SAR interpretation attributes (tone, texture, size, shape, pattern and context). During the last phase the influence of sensor and target parameters were considered on the information extraction process. The investigation has allowed to produce a semi-detailed map of the urban land use and land cover units for the Manaus area. This research has also shown that tone and texture were the dominant image attributes in the interpretation process, controlled by target (mainly surface roughness and moisture) and sensor parameters (lookazimuth, incidence angle and spatial resolution) which expresses distinct backscattering responses from distinct thematic classes. This investigation emphasizes the practical importance of the use of information derived from RADARSAT data for urban land use and land cover mapping in this complex and difficult environment.

ASSUNTO(S)

urbano sensoriamento remoto mapeamento uso da terra radar imageador manaus (am)

ACESSO AO ARTIGO

Documentos Relacionados