Research on Underwater Scene Reconstruction for Mobile Platforms Based on Rotating Scanning Sonar

High-precision underwater perception and scene reconstruction are critical techniques for marine surveying and resource exploration. Multi-sensor data fusion is currently the dominant method in underwater sensing. In this paper, a new approach for underwater sensing based on an integration of a 3D rotating scanning imaging sonar, an RTK (Real-Time Kinematic), and an IMU (Inertial Measurement Unit) systems onboard an unmanned surface vehicle (USV) is raised. By employing multi-sensor data fusion and image correlation calibration, combined with multi-view acoustic image synthesis, the system achieves accurate reconstruction of both water column and seabed scenes. The new system offers high reconstruction accuracy, and provides a cost-effective solution for scene reconstruction with a low requirement of the precise motion control of the USV platform. High-precision seabed imaging results have been validated through lake bed imaging tests.