Research on Algorithms for Edge Detection and Contour Tracking of Marine Robots

This paper uses video image analysis and preprocessing methods to study and analyze the environmental information collected by the Marine robot vision, and proposes the application of high-order fractional differential equation to the data information processing of the Marine robot vision algorithm. The object recognition algorithm of the Marine robot uses the processed visual images to process the object feature information. To obtain good working environment video image information for robot visual recognition, and provide reference for the next trajectory planning. Aiming at the problem of Marine robots recognizing environmental feature information, corner feature information and edge contour feature information of objects in video images are studied, Moravec corner detection algorithm is studied and analyzed, Harris corner detection algorithm and Canny edge detection algorithm are improved, and the improved corner and edge detection algorithm has very good robustness. Very good corner feature information and edge feature information are obtained for Marine robots.