The Research of Transmission Line Foreign Body Detection Based on Motion Compensation
Shengxi Jiao, Haiyang Wang · 2016
Foreign body has been identified as a significant problem in power transmission due to the potential for system damage. In this paper, the euclidean distance method and forecasting of drift of area are fused to remedy the defect of key frames redundancy. An six parameters model is based on K-ORB features, which is generated by applying K-means algorithm to cluster ORB(Oriented FAST and Rotated BRIEF) features. Then the target areas are obtained by three frames difference, which can be filtered by abnormal ratio analysis for eliminating interference area. Eventually, the movement locus of foreign body is drawn by gravity coordinates of target. Experimental results show that foreign body can be accurately and timely identified by the algorithm proposed in this paper.