Investigating the Detection of Overhead Ground Wire Broken Strand though Power Line Robots

Power line inspection and maintenance robots have been developed for more than two decades as a substitute for linemen to work under the dangerous working condition. The fault and obstacle detection through the information collected by robot sensors is an important research filed. Currently, visual information remains the most common, practical and easily collected data for robots amongst all kinds of sensing data. This paper presents a visual detection method for the broken strand fault. The broken strand fault can be detected by judging the intensity distribution of histogram of edge gradient(HOEG) in the following steps: at first, the region of interest (ROI) of captured images is selected; then, the HOEG descriptor is extracted in ROI as image feature; at last, the intensity distribution of HOEG descriptor can differentiate the broken strand fault from the normal wire. In addition, the location of detected broken strand fault is given for the practical repair. Both laboratory and field experiments results have demonstrated the effectiveness of the proposed method.