An Underwater Target Perception Framework for Underwater Operation Scene

摘要

This paper proposes an underwater target perception framework to comprehensively explore target information in underwater scenes, to improve the work efficiency and safety of underwater operations. This framework adopts a layered processing mechanism including water column imaging, constant false alarm rate detection (CFAR) detection, and local feature analysis, to accurately distinguish between false targets, static targets, and dynamic targets in the underwater scene, and obtain the motion trajectory of dynamic targets. The experiment is designed to simulate the underwater operation scene, and the results prove the effectiveness of the proposed framework.