A vision-based method for the broken spacer detection

Power line inspection is essential to the smooth running of the power grid. In the past, the power line inspection was mainly carried out manually, which means the liners had to inspect the power line in the field or watch the inspection video taken by UAVs or inspection robots. The manual inspection method is with disadvantages such as long time consumption and high manual labor cost. With the rapid growing of power grid, the demand for the power inspection has been continuously increased, but the manual inspection method can hardly meet the requirements for the power inspection due to its disadvantages. This paper presents a computer vision-based method for the broken spacer detection. The method is mainly implemented in three steps. First of all, the spacer is recognized in the region of interest. Secondly, the image morphology is used to extract the image feature. At last, we determine the broken spacer fault by the analysis of the connected domain in the image. Experimental results have successfully demonstrated the effectiveness of the proposed method.