Surface defect detection algorithm of high temperature casting slab based on improved YOLOv5s

Abstract

In order to improve the accuracy of surface defect detection of high temperature casting slab, an improved YOLOv5s surface defect detection algorithm is proposed. Firstly, Swin Transformer network structure is added to enhance the ability of feature extraction.Secondly, a coordinate attention mechanism is introduced to increase the sensitivity of position and direction information. Finally, a target detection layer is added to better realize feature fusion and enhance the generalization ability of the network. The improved algorithm has performed ablation experiments on the data set, which shows the effectiveness of the algorithm.

Keywords

• casting slab
• surface defect
• detection
• deep learning
• YOLOv5s