Lightweight Barcode Positioning Algorithm Based on YOLO Model

Abstract

With the advancement of Internet of Things (IoT) technology, barcode automatic identification systems have played a crucial role. An improved YOLO-MCG barcode localization algorithm was proposed to address the problems of interference, inefficiency, and poor real-time performance encountered by traditional barcode detection methods in complex backgrounds and field environments. First, to reduce the number of parameters and computational complexity, the MobileNetv3-small network is used to replace the backbone network. Second, a Convolutional Attention Mechanism Module (CBAM) is introduced to enhance the perception ability of target information and improve the detection performance of the model. In addition, Generalized Intersection over Union (GIOU) and Focal Loss are used as loss functions to enhance the localization precision of the model. The experiments show that the improved model achieves an mAP@0.5 of 97.8%, Precision of 96.4%, and Recall of 93.9%. The amounts of parameters are reduced to 35% of Yolov8, the amount of computation load is reduced to 25% of Yolov8, and the FPS is 105. The improved model can be used on resource-constrained mobile devices while meeting real-time requirements.