Estudo sobre a detecção de alvos em imagens SAR

Abstract

The use of synthetic aperture radars (SAR) has been growing in the last few decades, appearing to be an excelent alternative to Earth’s remote sensing and with applications varying from space exploration to change detection applied in the military field. This work is a study about SAR images and technics to extract results from them. The SAR images that are tested in this study are part of the swedish database CARABAS-II, with this work aiming to develop an algorithm capable of model and detect changes in those images. The problem of detecting targets is mathematicaly formulated, as well as the statistics adopted to model the images from the database. One test statistics is presented to employ the change detection method by a bivariate Rayleigh distribution. Metodologys for the target detection are developed, using a fix and adaptative value for the estimation of the target. Later, image processing technics are presented to remove false alarms