The Research On Obstacle Recognition Method For High Voltage Transmission Line De-icing Robot

Ice covering on high voltage transmission line is the world’s common problem faced by many national grid, many countries all have had transmission line accidents caused by ice covering, brings huge economic losses. The design of de-icing robots has become an important research direction to solve the problem of ice on transmission line.Because of the complexity of the transmission line environment, a variety of obstacles on the transmission line should be accurately identified for robot working stably and efficiently on-line. Extracting features from images caught by the armed cameras on robot is the key to improve robot autonomy, and also is the premise of robot behavior planning.Firstly, the ice covering seriously effect on power transmission is introduced. Then, the research and development of inspection and de-icing robot on transmission line and the navigation method for transmission line robot are summarized. On the basis of the analysis for structural characteristics of transmission line, three robot structures and there de-icing schemes are designed, and then various line components and the problem need to be resolved in the process of robot autonomously working are described.Secondly, the de-icing robot obstacle recognition method based on shape features is introduced. To improved the real time of the method, images are processed in Regions of Interest (ROI); method of image segmentation in HSI color model in succession using Otsu method to select the threshold is used to segment transmission line and background; the linear feature is applied to transmission line recognition and orientation, progressive probabilistic Hough transform (PPHT) based on image edge gradient is used to judge line segments endpoints precisely; the circular and elliptical features are applied to insulator recognition, image edge gradient information is used to reduce the computation in Hough transform for circle detection.Thirdly, obstacle recognition method based on image local invariant features for template matching is put forward. First, image SURF features are extracted, and the two-threshold sequential algorithm scheme is applied for features clustering, then one k-d search tree is established for each cluster and BBF priority search method is used to match template features, finally, the RANSAC robust estimation algorithm is presented to eliminate the error of matched points and estimate the homography of template and image, and then identify obstacles exactly.Finally, the visual surveillance and remote control system of de-icing robot is introduced; robot partial independent and remote human-computer interaction mode and de-icing robot obstacle recognition method based on human-computer interaction is put forward.