Method and device for detecting surface defects of cotton covered wire

The invention relates to the technical field of target detection of deep learning, and particularly provides a cotton covered wire surface defect detection method and device, and the method comprises the steps: collecting corresponding cable defect data from a production line, and carrying out the preprocessing; a YOLO v5 network structure is established, a neck network is improved, and an original PANet network is changed into a BiFPN network; nWDloss is fused in the original YOLO v5, so that the small target detection capability is improved; adding a CBAM attention mechanism; adding a prediction head for small target detection; and finally, sending cable surface defect data into the improved YOLO v5 network model for training to obtain a target detection optimal model and training number result data, and evaluating a model detection effect according to the final accuracy rate, the recall rate and the finally obtained MAP value. Compared with the prior art, the invention can improve the efficiency of inspect