文章基本信息

标题：EXTRAÇÃO DE RODOVIAS EM IMAGENS DIGITAIS DE ALTA RESOLUÇÃO COM O USO DA TEORIA DE MORFOLOGIA MATEMÁTICA
本地全文：下载
作者：Aline Sayuri Ishikawa ; Erivaldo Antonio da Silva ; Rodrigo Affonso de Albuquerque Nóbrega 等
期刊名称：Revista Brasileira de Cartografia
印刷版ISSN：0560-4613
电子版ISSN：1808-0936
出版年度：2011
卷号：63
期号：01
出版社：Sociedade Brasileira Cartografia - Geodesia
摘要：This paper proposes a method for road extraction from high resolution imagery based on Mathematical Morphology theory. The proposed method consisted on the development of mathematical morphology and structuring elements based routine using the SDC Morphology Toolbox for MATLAB. The routine uses operators as opening, closing, skeletonisation, erosion, dilation, morphologic gradient, among others. Prior to processing, the whole study area was subset into 11 sub-images due to the limitation of MATLAB and available hardware on computing large raster data. Thus, the developed morphologic routine was capable to be replicable for all 11 sub-images by considering only changes for the thresholds values, which differ for each image. The 11 sub-images were mosaicked later than the road extraction to recompose the overall test area. Next, the road vectors were generated to compute the quality analysis of the result obtained with the extraction. The assessment of the result was based on linear correspondence analysis. The high degree of the completeness metric indicated quite satisfactory results. It proves the effectiveness of the Mathematical Morphology theory as part of the features extraction process.
关键词：Keywords: Mathematical Morphology, Road Extraction, Correspondence Analysis.