Radar image ship target detection method considering global and local information perception

• Main Authors: WANG RUFEI, HUANG YULIN, PEI JIFANG, YANG JIANYU, XU HONGLIN, HUO WEIBO, ZHANG YIN, YANG HAIGUANG
• Format: Patent
• Language: Chinese; English
• Published: 19.09.2023
• Subjects: CALCULATING; COMPUTING; COUNTING; PHYSICS

The invention discloses a radar image ship target detection method considering global and local information perception. The method comprises the following steps: S1, obtaining a radar image sample; s2, identifying an ocean area and a land area in the radar image by adopting a self-adaptive threshold method; s3, extracting a stable extremal region in the image by adopting a maximum stable extremal region method, and taking the stable extremal region as a candidate ship region; s4, performing superpixel segmentation on the whole image by using a simple linear iterative clustering method; s5, saliency value calculation is carried out on all the candidate ship areas, and when saliency values meet preset conditions, the candidate ship areas are judged to be final correct detection areas. The global and local information perception of the radar image is comprehensively considered. On one hand, context information in a scene is introduced into ship detection in a complex scene, and more effective information is prov