基于YOLO_v3的UL标签识别方案

UL Label Recognition Algorithm Based on YOLO_v3

由于PCB电路板中UL标签的不规范标记与使用,给PCB印刷厂商造成监管困难的问题。因此,提出了一种基于YOLO_v3算法的UL标签检测方案。方案通过收集并制作了含有UL标签的图像数据集,对训练样本进行变化扩充后,使用YOLO_v3网络进行训练与检测。实验表明,该方法在复杂且干扰性强的PCB内UL标签检测过程具有较强的鲁棒性和有效性,可以精确且快速地完成UL标签的检测。

Due to the non-standard marking and using of UL label in PCB circuit board,it is difficult to supervise PCB in printing manufacturers.Therefore,this paper proposes a UL target detection method based on YOLO_v3 algorithm.In this method,the image data set containing UL tag is collected and made.After changing and expanding the training samples,Yolo is used_ V3 network for training and detecting.Experiments show that this method has strong robustness and effectiveness in the detection process of UL tag in complex and highly disturbing PCB.It also can complete the detection of UL tag accurately and quickly.