In recent years,with the rapid development and extension of grid structures at all levels of the power system,the coverage of overhead line corridors is becoming closer and closer,and the electrical accidents caused by the damage of transmission lines by external forces are becoming more and more frequent,which has become an important problem to be solved urgently in the safe operation of power system.In order to realize the effective early warning of external force damage prevention of overhead transmission line,it is of great significance to develop an anti external force damage system of overhead transmission line with the functions of video capture,image analysis,independent judgment and early warning prompt.This paper takes the anti external force damage of overhead transmission line as the research object.According to the degree of foreign object harm,it is divided into intrusion existence and damage threat,carries out the research on the relevant algorithm of the anti external force damage system of overhead transmission line,completes the design and development of software system and hardware system combined with the algorithm.The main research contents and result of this paper include:(1)This paper discusses the overall structure principle of the anti external force damage system of overhead transmission line,including the design of image acquisition radar touch alarm linkage and on-site warning function,the background B/S structure of the monitoring center and the composition structure of the front-end monitoring extension.Analyzes the working principle of the front-end video recording device and the working principle of the background early warning software system respectively,complete the hardware design of the anti external force damage system of overhead transmission line.(2)The image processing algorithm of preventing foreign object intrusion of overhead transmission line is studied,and the foreign object intrusion detection algorithm of overhead transmission line based on image difference analysis is proposed.Taking a group of human target detection under tower as an example,the effectiveness of the foreign object intrusion detection algorithm of overhead transmission line based on image difference analysis is verified through the process of extraction,filtering and denoising,image gray correction,image edge enhancement and feature extraction;(3)This paper studies the image processing algorithm of anti external force damage of overhead transmission line,compares and analyzes different deep learning target detection algorithms,selects R-FCN model detection algorithm according to the image characteristics of overhead transmission line.Firstly,the model components of R-FCN are selected,which include feature extraction network,network depth and candidate frame search algorithm.Res Net101 is selected as the feature extraction network and RPN module is selected as the candidate frame selection algorithm through experimental analysis.Aiming at the over fitting problem of the traditional R-FCN model in sample training,the transfer learning principle and softening non maximum suppression algorithm are introduced to improve the R-FCN model.The improved R-FCN algorithm is compared with other target detection algorithms,experimental results demonstrate that the m AP30,m AP50 and AR metrics of the improved RFCN model are significantly improved.Finally,experiments on four target detection models using test sets on Pascal VOC 2007 and Pascal VOC 2012 show that the improved R-FCN network has better detection results.(4)This paper studies the effectiveness of relevant algorithms of anti external force damage system,select smoke,large vehicles and floating objects in the air for experimental verification against the invasion of external objects of transmission line;Aiming at the external force damage of transmission line,large crane is selected as the target for experimental verification,which effectively proves the effectiveness of the improved R-FCN algorithm proposed in this paper.Complete the software platform development of the anti external force damage system of overhead transmission line.The anti external force damage system of overhead transmission line designed in this study can achieve effective early warning of potential damage by external forces,and the relevant algorithm can provide a reference for the design of such systems. |