Region-based airplane detection in remotely sensed imagery

This paper studies on airplane target detection in remotely sensed imagery. The procedure of airplane detection consists of two steps. First, segmentation is performed on the original image to create discrete image regions. The image regions are used to determine which region belongs to candidates of target area. Second, each candidate area is re-processed into a binary image. In the binary image, the foreground is taken for the candidate target object. A simple statistic is computed from the binary image for target detection. Experiments with high-resolution remote sensing images show that the regularly visible airplane targets could be detected successfully.