Development of a System for Automatic Identification of a Problem Areas of Extended Objects Using AUV

The paper describes the developed system for automatic identification of problem areas of extended objects using autonomous unmanned underwater vehicles (AUVs) based on data obtained from technical vision systems (TVS), in particular multibeam sonars. The proposed system allows a three-dimensional model of the pipeline to be dynamically constructed in real time on the AUV's on-board computer, based on point clouds received from the TVS. Based on this model, the position of an extended object in space and its curvature are automatically determined, as well as the degree of immersion of the pipe in the seabed. The system generates information about the condition of the pipeline, which can be sent to the post using a hydroacoustic communication channel. The software implementation of the system is implemented in the Python programming language. The system has been studied using numerical modelling in the CoppeliaSim environment. The simulation results confirmed the operability of the system and the high efficiency of its use.