Inspection Robot for Submarine Pipeline Based on Machine Vision

Abstract In order to improve the intelligence of submarine pipeline inspection, reduce manpower and material resources, and improve inspection efficiency, an underwater pipeline inspection robot based on machine vision is proposed. This paper introduces the existing pipeline inspection methods and technologies, and analyzes the feasibility of pipeline inspection and inspection technologies based on image processing. On the basis of this subject, the corresponding circuit and embedded system are designed, and the related images are sampled, processed and analyzed, and the images are subjected to threshold segmentation, binarization, edge corrosion and other operations. Furthermore, the experimental site and underwater robot were built, and the pipeline defects were detected in real time by using the recognition method based on neural network.