Abstract:

Because of deicing robot works in icing power lines, identifying obstacles exist some shortages, such as the difficult to distinguish various types of obstacles, low accuracy, and so on. To improve the recognition ability of robot, a kind of adaptive threshold wavelet transform edge detection algorithm to extract the edge of obstacle is designed. And according to the structural characteristics of power line obstacles, an effective method for eliminating partial interference background based on power line position constraint is designed in the process of obstacle edge extraction. The wavelet moment is introduced, the wavelet moment of the edge image is extracted as the feature matching data of the obstacle. According to the principle of neural network and particle swarm optimization algorithm, a wavelet neural network method based on particle swarm optimization is proposed for obstacle recognition and classification. The particle swarm algorithm is used to replace the traditional gradient descent method, the inertia weight factor is improved and the parameters of wavelet network are optimized. The experimental results show that the obstacles such as counterweight, suspension clamp and strain clamp on the power line can be effectively recognized by the proposed method, and the recognition accuracy is higher than the conventional recognition method.

Key words: obstacle recognition, particle swarm optimization algorithm, wavelet moment, wavelet neural network, deicing robot