Change detection by masking reversible areas

Many issues addressed today in remote sensing based on data from multi-temporal satellite images are related to change detection, for the implementation techniques having aims at the location, characterization and quantification changes in the state of an object in same scene area which have evolved between different instants. Unfortunately, most of these techniques confront the problem of persistent shadows and cloud cover, even after selecting the best images. Knowing, that clouds and shadow are considered as false alarms (no real change). However, in this paper, we propose to detect the changes by eliminating the false alarms (shadow and clouds) from the areas covered by the shadows of the objects and the cloud cover. In order, to improve and obtain almost real change results. For this we used the combination of two methods: the change detection method, clouds and shadows detection method, to eliminate their information detected as change. The obtained results show the efficiency of the proposed method.