Abstract:When a traditional helicopter is used to inspect an overhead transmission line, the pod or cradle head equipped with testing equipment must be manually controlled so that the target to be tested is always within the range of vision, which requires highly concentrated observation of video by the inspectors. In order to reduce the burden on the lineman, an automatic tracking method for the photoelectric pod is proposed. It selects Struck algorithm and uses Haar features to describe moving targets. In the tracking process, it searches the adjacent location of the target region of the previous frame, and uses structured output SVM to deal with the selected regions for binary classification, the candidate region with the highest score is obtained as the result of tracking the current moving target. The experimental results show that the average deviation rate (field Angle) of the algorithm is 1.6%, and the target loss rate is 14%, which can track the key parts of transmission line, such as tower, insulator and equalizing ring, in real time. |