Study on the algorithm for automatic plane classification from remote sensing images with mid-high resolution

A novel algorithm for automatic plane classification which is adapted to aeronautics and astronautics remote sensing images with mid-high resolution is proposed. Plenty of prior knowledge is used in the process of the segmentation and the classification. According to improved regional segmentation algorithm (IRSA), the threshold value is chosen accurately and automatically during the regional segmentation. The binary tree classifier is designed and applied to the plane automatic classification. Using simple geometric features which are extracted from the plane object, good effect and efficiency of classification are achieved. Several experiments to verify the proposed method are given with 100% detection rate, i.e. 0% false alarm rate and 0% miss detection rate.