Enhanced defect sensing technology in turbid water environments using multi-beam sonar

Abstract

In this paper, we report a novel defect perception technology utilizing multi-beam sonar for applications in turbid water environments. Our goal is to improve the precision and speed of identifying target image defects. We categorize the target image recognition dataset following specific guidelines and devise a target image imaging model customized for the distinct characteristics of turbid water settings. By employing the weighted time average (WMT) algorithm, we calculate the time window for each beam within the water environment. Moreover, we utilize the phase difference sequence method to enhance target image details in turbid water, and leverage the time of arrival (TOA) estimation method to suppress background noise and sidelobes. Through the implementation of a dynamic detection threshold, our technology facilitates defect perception in turbid water environments using multi-beam sonar. Experimental results demonstrate that this method achieves an accuracy of 96.05 % in recognizing image defects in turbid water environments, significantly enhancing both the accuracy and efficiency of defect recognition. It also overcomes the typical challenges of underwater detection in turbid and low-light conditions.