The Detection Method Of Underwear Ejection Target Based On Multi-beam Sonar

Underwater eruption targets are widely found in a variety of marine activities and are closely related to the exploitation of marine resources and the protection of marine environment.This paper studies the underwater ejection target detection technology based on multi-beam sonar aiming at the demand of underwater ejection target detection to realize real-time detection of underwater ejecting targets,which is under the support of the national science and technology major projects "Key technology of real-time monitoring of sea tube leakage based on AUV platform".Firstly,this paper gives a brief overview of domestic and foreign underwater ejection target detection technology and introduces the underwater ejection target detection technology based on multi-beam sonar.Secondly,basing on the shortcomings of the existing multi-beam sonar underwater ejection target detection technology,a new underwater ejection target detection scheme is proposed.This new scheme utilizes beam-based detection algorithm and image-based detection and identification algorithm,and then transplants the algorithm to the DSP data processing platform.The multi-detect algorithm uses the VI-CFAR detector to pre-detect the beam signal and then uses the bottom detection algorithm to determine the region where the suspected underwater eruption target is located.The image processing algorithm studied local water body imaging technology,image preprocessing technology,image segmentation technology and dynamic feature extraction technology,thus supporting vector machine to realize the identification of underwater eruption targets.Finally,this paper completes the transplantation and real-time implementation of the algorithm based on the DSP data processing platform.Through the processing and analysis of the field test data,the paper verifies the effectiveness of the algorithm for underwater ejection target detection.By testing each module of DSP data processing platform,the feasibility and effectiveness of the real-time detection scheme of underwater eruption targets are verified.