VISUAL NAVIGATION METHOD FOR ARRANGING INSPECTION OF UNDERWATER PIPELINE BY AUTONOMOUS ROBOT

The article reviews a simulation modeling solution to the problem of automatic inspection of underwater pipelines using autonomous underwater robots. The proposed method for determining the position of the robot in relation to the inspected pipeline in the immediate vicinity of the object is described. The method is based on the processing of video information and measurements of a linear laser using the method of visual navigation. Original pipeline identification algorithms are proposed for use with both mono and stereo cameras. Functional testing of the developed algorithms and performance evaluation was carried out on virtual scenes with in the framework of the simulation modeling complex developed by the authors, distributed on a multiprocessor computing architecture. Computational experiments have shown that the accuracy of the calculated parameters of the relative position of the autonomous underwater robot and the pipeline meets the requirements of the autonomous underwater robot control system.