Electric power overhead conductor detection method based on course learning strategy

The invention discloses a course learning strategy-based overhead power conductor detection method, which comprises the following steps of: performing single-sample multi-target frame labeling on a data set, performing segmented labeling and segmented detection on an overhead conductor, and marking the trend of the overhead conductor; image complexity calculation based on wavelet-entropy is carried out on the data set pictures, the data set is divided into a simple data set and a complex data set, the two data sets are successively sent to a YOLOv5 network for training and learning, and the overhead conductor detection precision is further improved; for an overhead conductor picture which still has error detection after YOLOv5 network detection, target frame processing is carried out by utilizing direction gradient statistics, an error detection frame is removed, a complete and accurate overhead conductor target detection frame is reserved, and a complete and accurate overhead conductor is reserved. Under the