Research On Visual Control System Of Insulator Washing Robot

Railway catenary insulator is exposed to the external environment for a long time.Affected by regional climate,environmental change,railway transportation conditions and other factors,it is very easy to form a pollution layer on its surface.Excessive insulator fouling may lead to "pollution flashover",which will affect the normal operation of railway catenary circuit and bring huge economic losses.Therefore,insulators need to be cleaned regularly.Because of its advantages of simple operation,low working cost and good cleaning effect,water washing truck is widely used in many railway administration group companies in China.However,the traditional water washing truck relies on manual operation and has a low degree of automation.With the rapid development of computer vision technology,intelligent robot based on vision control has been widely used in various fields,and the control accuracy and efficiency have been greatly improved compared with manual operation.Therefore,visual control technology has become a hot research topic in the development of insulator washing automation.In order to realize automatic water washing of insulators,a control system of automatic water washing of insulators based on computer vision is proposed.The main research contents of this thesis are as follows:1)The actual working environment of the insulator water washing robot is analyzed,the working process of the washing robot and the insulator automatic washing scheme controlled by computer vision system are formulated,and the main components of the insulator water washing robot system are determined.2)After analyzing the on-site installation environment and image characteristics of insulators,the detection algorithm based on YOLOv4 was selected to identify insulators in complex background,and the transfer learning strategy was used to improve the training efficiency of the model.Through experimental comparison,an appropriate fusion strategy is selected and the improved KCF target tracking algorithm is combined to continuously track the insulator target,so as to meet the requirements of real-time and accuracy.3)According to the requirement of matching speed and accuracy,insulator target identified by detection tracking algorithm is extracted and matched by using corresponding point extraction and matching technology in binocular vision,combining with improved ORB feature extraction matching algorithm based on image pyramid scale to extract and match insulator corresponding feature points of left and right cameras.4)Based on the camera imaging model and calibration method,the plane target is used to calibrate the internal and external parameters of the binocular vision system,the mapping relationship between the world coordinate system and the image coordinate system is established,and the three-dimensional coordinates of the insulator relative to the camera are restored through the binocular vision measurement model.5)According to the requirements of aiming stability,accuracy and rapidity of the insulator washing robot,the image coordinate deviation processed by the visual algorithm is converted into the angle deviation of the washing water gun by using the image-based feedback control method,and finally converted into the servo motor control signal to adjust the attitude of the washing water gun and binocular camera to aim the insulator.Finally,according to the characteristics of visual servo model of insulator washing robot,the traditional PID algorithm is improved to raise the dynamic response ability of washing device and provide algorithm basis for automatic washing aiming.6)The experimental platform of binocular water washing robot is designed and built,and the automatic water washing experiment of insulator is carried out to verify the feasibility of the method.