Research on Laser Intelligent Removal Foreign Body in Ultra-long Distance Power Grid

Abstract

The remote power grid foreign body remover based on laser technology has been widely used in the production work currently, however the manual operation of power grid laser remover to remove foreign bodies generally has the difficulties of high control difficulty and low precision. Therefore, a foreign body detection and laser clearance method based on Mask R-CNN instance segmentation is proposed. A more accurate method for detecting the category and position of foreign body targets was proposed by constructing the database of electrified wire netting and foreign body winding shape, filtering analysis and feature extraction were carried out on the foreign body images hanging on the wire, though which can track and aim the characteristics of foreign body and give the optimal laser cutting path. The experiment result shows that the success rate and remote recognition accuracy of this method both work well, which indicates that this method has a certain practical value in the identification and removal of foreign bodies in ultra-long distance power grids.