Desenvolvimento de uma ferramenta web no processo de segmentação de imagens

Abstract

Remote sensing is a set of methods for capturing information from terrestrial phenomena and features through sensors, without direct contact, and the methodologies and techniques for storing, processing, analyzing and interpreting this information are indispensable. This work details a method, available through an online tool with client/server graphical interface, which decomposes an image of spectral bands into a column vector. We show the formal mathematical model of the Fuzzy ckMeans algorithm and we propose the Fuzzy ckMeansImage. Still, we show the overlay process for grouping an image given a number of clusters. In order to validate the proposal, the algorithm was applied to a satellite image, referring to an area used for rice cultivation located on the western border of the state of Rio Grande do Sul. Clustering results are discussed by a specialist in remote sensing.