Application of Joint Immune Ant Colony Algorithm and Fuzzy Neural Network to Path Planning and Visual Image Processing of Inspection Robot in Substation

Abstract

There are kinds of power equipment and complex structure in high voltage substation, which need to be inspected regularly. At present, substation inspection robot is widely used in process of substation inspection. It mainly inspects the insulator, grading ring and other external accessories of high-voltage power equipment. In the process of inspection, the robot is equipped with infrared imager and ultraviolet imager to store the photon distribution map and the temperature distribution map of high-voltage power equipment, and establish ultraviolet and infrared image database of the equipment. In the process of inspection, the robot should avoid area with higher electric field intensity, and traverse the typical high-voltage electrical equipment in the substation in shortest distance, and the robot arm can effectively carry out image multi-angle acquisition. On the other hand, in the process of image acquisition, the robot should realize the image region segmentation, focusing and photon acquisition operation of the insulator and grading ring of the equipment, and apply the deep learning method to realize the automatic matching and comparison between the collected image and the original image library, to distinguish the operation state of the high-voltage power equipment and find the latent insulation fault. Among them, shortest path of low field strength and path of traversing equipment are automatically realized by the immune ant colony algorithm, and the image database deep learning and automatic matching module are realized by the fuzzy neural network. In this paper, immune ant colony algorithm and fuzzy neural network are applied to the path planning and visual image post-processing of substation inspection robot, which has certain theoretical and engineering guiding value for intelligent inspection of substation high-voltage power equipment.