A Method of Segmenting Texture of Targets in Remote Sensing Images Based on Granular Computing

In order to solve the aircraft targets segmentation and recognition problem with background interference in remote sensing images, a base-on granular computing method to division texture is used in this paper. By granulating remote sensing images, the gray level co-occurrence matrix and Gabor filter can catch these characters in elementary granules. And these elementary granules are incorporated to a result of texture division in quotient space. For illustration, a remote sensing image with two aircraft targets is utilized to show these steps of segmentation. Empirical results show that the decision rule with highest confidence measure will be used as the final decision rules in the MADA problem with multiple conflicting decision attributes. The base-on granular computing texture characters segmentation method can effectively solve the aircraft targets segmentation and recognition problem with background interference in remote sensing images are solved.