Research on surface defect detection of disc harrow based on YOLOv5

In view of the problems of low efficiency, missing detection and error detection, and inability to meet the requirements of intelligent manufacturing in the way of manual detection of the surface defects of agricultural machinery rake, in order to use computer vision technology based on depth learning to replace manual detection of the surface defects of agricultural machinery rake, this paper proposes a surface defect detection algorithm based on YOLOv5. Firstly, CA attention mechanism is introduced to make the network pay more attention to important channel information and enhance the ability of feature extraction network to extract effective features; The CIoU NMS designed based on the generalized intersection ratio CIoU is used to replace the traditional non-maximum value in the original algorithm to suppress NMS and reduce the problem of missing detection; In combination with SIoU Loss and FocalL1 Loss, Focal-SIoU Loss is proposed as a bounding box loss function to reduce the negative impact of low-quality samples, accelerate the convergence speed of the model, and improve the regression effect of the model. The experimental results show that the improved model proposed in this paper has better performance and can improve the detection efficiency compared with the baseline model.