Abstract: Due to the stringing construction damage and the long-term effects of the external environment, untwisted strands and even broken strands are prone to occur in overhead transmission lines. Timely diagnosis and overhaul of the defects are therefore very significant for the power grid security. Inspection by robot equipped with visual detection devices is an effective method in modern condition-based maintenance of power transmission lines. On the basis of analyzing the constraints and the distribution of transmission line textures, an untwisted strand judging model is built based on the texture slope distribution and an automatic diagnosis method of the untwisted strands of overhead transmission lines is proposed based on digital image processing. This method gives the results of image diagnosis by using the judging model after calculating the texture characteristic distribution of the transmission lines which are extracted from the image taken by the robot. Experiment shows that this method can quickly detect the fault information of the untwisted strands, and satisfy the real-time requirement of online inspection.

Key words: inspection robot, visual image, texture characteristics, untwisted strand