Distance image segmentation method for terminal sensitive missile linear array laser radar

Abstract

In order to enhance the detection and recognition performance of the terminal sensitive missile on the ground armor target in different scenarios, the application background of the terminal sensitive missile-loaded linear array laser radar was fully considered, and the integrated cross-scan line method and the ground-distance image point of the gradient connected domain were proposed. Cloud segmentation algorithm was used to improve the segmentation effect on the ground and target image. Firstly, the original distance information obtained by laser radar scanning was converted into the horizontal ground height value, and the spatial slope was transformed into the horizontal plane by the cross-scanning line method to enhance the adaptability on different terrains. Then, the ground point cloud connected domain algorithm was used to extract the ground. Point cloud and morphological gradient threshold method were used to segment the target point cloud. Finally, the geometric similarity of the feature segmentation effect was calculated. The experimental results show that the proposed algorithm has good applicability to various terrains such as positive slope and side slope. It can accurately and effectively segment the point cloud in different heights, terrains and slopes, and improve the target recognition performance of terminal sensitive missile.