Détection des changements dans les images satellitaires

Abstract

Change detection is an integral part of the analysis of satellite imagery, and it has been studied for several decades. It consists of comparing a registered pair of images of the same region and identifying the parts where a change has occurred, it allow to follow the evolution over time of a region of interest through technical changes upon detection So these images are a tool of choice in the management of natural resources. This requires a methodological approach appropriate image processing to the use of such data. In this work we compared the performance of three methods (Difference, ACP-Kmeans, Logmean) for the detection of changes in satellite images using the different evaluation metrics (FA, DR, Kappa, Precision, Recall, TC), and we were found that the ACP-Kmeans method gives a binary DC mask with better precision compared to other method.

Change detection is an integral part of the analysis of satellite imagery, and it has been studied for several decades. It consists of comparing a registered pair of images of the same region and identifying the parts where a change has occurred, it allow to follow the evolution over time of a region of interest through technical changes upon detection So these images are a tool of choice in the management of natural resources. This requires a methodological approach appropriate image processing to the use of such data. In this work we compared the performance of three methods (Difference, ACP-Kmeans, Logmean) for the detection of changes in satellite images using the different evaluation metrics (FA, DR, Kappa, Precision, Recall, TC), and we were found that the ACP-Kmeans method gives a binary DC mask with better precision compared to other method.