Abstract:

Change detection at the pixellevel is use the original gray information of different temporal remote sensing images, which are registered previously, to find the change information of the environment. It is perspicuous and easytorealize, and is a generally applied change detection method. As far as the actual algorithm concerned, it can not work allaround among adaptability, robustness, accuracy, and timeeconomy. So in this paper, a novel fusion change detection method based on the DS evidential theory is introduced, which combines the results from different change detection algorithms by the DS fusion tactics and rules, and then the changes that are concerned can be detected at the decisionlevel. The results of the experiments show that the algorithm improves the accuracy and robustness of the change detection, reduces the choosing bound of the threshold, and decreases the impact of the threshold altering on the algorithm.

Key words: change detection;remote sensing image;pixellevel;DS evidential theory;fusion