Aplicação de morfologia matemática e equações diferenciais na extração de malha viária em imagens digitais

Abstract

A country that has the continental dimensions such as Brazil needs to have a solid cartographic base to take important decisions about urban planning and consequenting to all national territory. To realize conventional surveys with the intention of minimize the outdated, the price would be a lot expensive. Thinking in unexpensive methods and fast, with the intention of minimize this outdated, the Remote Sensing Images are contributing to minimize this problem. Amongst many tools of Digital Processing Images - DPI, were used in this paper the partial differential equations (PDEs) were used in the smoothing phase of the images, since the model proposed by Barcelos and Chen softens an image without compromising the contours of the features contained in the scene. Afterwards, the morphological operators were used with the purpose of eliminating undesirable features and extracting the features of interest. Mathematical Morphology has as objective to improve visually the usually the original image quality and to extract the relevant information. Considering that Morphological operators use grays level as the major recognition pattern, the proposal of this research is to apply previously in the segmentation process. Mathematical Morphology tools are connected in the Matlab software. The extracted features can be used in conventional processes of updating cartographic products. The obtained results are positives and showed the use potential of mathematical morphology theory in Cartography.