Study On Visual Identification Method Of Obstacle On High Voltage Transmission Line

Now,people's lives and the country's development have been completely inseparable from the power,which requires the country's basic facilities to include a very reliable power supply system.High voltage transmission line as the main remote power transmission mechanism,it is a basic requirement is safe and stable.However,the actual working environment of high-voltage transmission lines put forward higher requirements,the transmission line often fail to work and the detection for this phenomenon is very difficult due to artificial limitations.The development of robots provides a very good solution to this problem,by using high-voltage line inspection robot walking on the line;it can determine whether there are potential problems according to the picture taken by the camera.However,some accessories on high-voltage lines,such as shock hammers,spacers,and clamps,and so on,will hinder the robot's walking,so a very urgent problem is to solve the identification of these accessories detection,providing obstacle signals for the robot.Aiming at the structure characteristic of 500 k V demonstration line,and the specific content and requirements of task,this paper proposes a visual detection and identification scheme for high voltage transmission line obstacle.In this paper,the characteristics of the operating environment and the types of obstructions are analyzed,and focus on the structural characteristics and installation methods of the obstructions such as the hammer,the spacer,and clamp.On this basis,the overall structure of the inspection robot is designed and the identification of obstacle target is analyzed,including the design requirements,the difficulty analysis and the design of overall recognition algorithm.Accurate and fast identification of obstacle targets is the prerequisite for the robot to cross the obstacle,which indicates the obstacle operation of the robot.In this paper,the recognition and detection algorithm of the target is designed and analyzed in detail.The algorithm mainly consists of two parts: feature extraction and classifier design.The total recognition algorithm is divided into two levels: coarse recognition and precise recognition.The coarse recognition is based on the Haar-Like feature,and the Adaboost is classifier.The precise recognition is based on the HOG feature,and the support vector machine(SVM)is used as the classifier.Based on the designed model,the experiment and analysis on the actual working line of the field high voltage power line are carried out.The results show that this two-stage screening method achieved a good classification result by applying to the high voltage line for the detection and recognition of the obstacle target.With the help of QT,a simple human-computer interaction interface is designed to display the target detection results in time.