Attention mechanism foreign fiber image recognition algorithm

Abstract

In order to solve the problem of poor detection effect of heterotropy fiber in cotton like cotton, a method of heterotropy fiber detection based on improved YOLOv5 was proposed. The CBAM module of attention mechanism was introduced to build the improved CBAM-YOLOV5 model. The real cotton fiber image data set was divided into training set and test set according to the ratio of 4:1, and six image augmentation methods such as translation and rotation were used to expand the data set. The YOLOv5 model before and after improvement was compared. The experimental results show that the improved YOLOv5 model can better identify the cotton-like fibers and improve the recognition accuracy by 7.68%.