Classification and Positioning of Circuit Board Components Based on Improved YOLOv5

Abstract

With the development of electronic technology, the amount of electronic solid waste is increasing. The waste circuit board has lost its original function, but some components have not yet reached the standard of scrapping, and can be dismantled by dismantling. Recycle it and use it on a new circuit board. Aiming at the problem of automatic dismantling and recycling of electronic components on circuit boards, a combined algorithm based on YOLOv5 and hierarchical classification algorithm is proposed. First, images of components with similar features and indistinguishable features are integrated into new category data, and a classification model is trained after cropping. Second, the output of the YOLOv5 network is subdivided into predicted categories by the classification network, and then the output of YOLOv5 is redistributed to complete the recognition. Finally, the algorithm is applied to the detection and identification of circuit board components. It can be seen from the experimental results that the combined algorithm has high recognition accuracy and accuracy in the classification and positioning of circuit board components. The algorithm effect is 38% higher than the original YOLOv5 model.