A reliability-prioritized adaptive real-time MBES data processing method for AUVs in complex environments

With their advanced autonomous navigation capabilities, AUVs are highly effective tools for detecting and analyzing intricate underwater structures. The challenges associated with real-time manual control place significant demands on AUV intelligence, particularly in data processing and environmental understanding. To enhance AUVs’ capacity for thorough exploration and identification of complex underwater structures, this research aims to develop an adaptive approach for processing MBES data. Conventional MBES data processing methods are largely designed for post-processing or real-time processing under predefined conditions. However, these methods are insufficient for addressing the unpredictable and complex scenarios encountered during AUV surveys. To overcome these limitations, this paper presents a novel strategy for MBES data processing. At its core is an adaptive algorithm designed to suppress unreliable MBES data, optimized for handling complex 3D targets, combined with a customized approach for target surface reconstruction. The proposed method significantly enhances AUVs’ ability to process MBES data and analyze underwater target surfaces in real time, thereby advancing their capabilities for autonomous exploration. • This study aims to enable AUVs to autonomously process MBES data during missions. • Adaptive MBES processing for AUVs operating in complex, unpredictable underwater environments. • The method derives reliable MBES data and rebuilds a clean target surface from noisy signals. • To verify the universality of our method, we prepared a variety of test data.