Algorithm Research And System Implementation Of Underwater Target Detection And Tracking Based On Forward-looking Sonar

With the transformation of the national marine strategy and the improvement of the marine status,the underwater acoustic technology of underwater submersibles has attracted much attention at home and abroad.Therefore,the study of underwater target detection and tracking based on forward-looking sonar has great practical significance for the development of underwater detection and tracking technology of underwater submersibles.This paper focuses on the underwater obstacle detection and target tracking problem based on forward-looking sonar.The image processing method based on forward-looking sonar image is studied,and a target detection and tracking system suitable for single-beam stepping forwardlooking sonar is realized..Based on the characteristics of Micorn DST forward-looking sonar images,this paper studies the nearest obstacle object detection algorithm based on forward-looking sonar images,with median filtering,fuzzy enhancement based on log membership degree,maximum entropy segmentation,suspected target screening and improvement.The inverse distance weighted interpolation backward mapping algorithm is composed.Then according to the actual needs of the cable underwater robot platform,an acoustic detection system based on forward-looking sonar is realized and the expected experimental results are obtained.After that,this paper analyzes the various grayscale,shape and texture features of the underwater target in the Micorn DST forwardlooking sonar image.Finally,we choose to use the invariant moment to construct the feature set as the observation model of the particle filter tracking method.And the classical particle filter algorithm using gray template matching as the observation model and the particle filter algorithm with feature set as the observation model are compared.Finally,a particle filter algorithm based on forward-looking sonar image is realized,and then the sound detection based on this paper is studied.The system completes the target tracking function of the sound detection system.The sound detection system realized in this paper can provide the early warning function of the underwater obstacles,has a complete software and hardware structure,and tests the performance of the sound detection system through pool experiments and ocean experiments.The implementation results show that the sound detection system can provide a stable and reliable near obstacle warning function for underwater robots.According to the sound detection system realized by the research,a particle tracking-based target tracking module is developed.After the actual data verification,the target tracking module can effectively track the underwater target and provide the target tracking function for the sound detection system.