基于机器视觉的快递面单识别

Recognition of waybill based on machine vision

为有效解决分拣环节人工投入成本高、后台信息更新速度慢和分拣效率低等问题,提出基于机器视觉同时对一维码和三段码定位识别方法。改进一维码定位方法,利用“线扫描”提高一维码解码正确率;改进Faster R-CNN目标检测方法,实现不同区域三段码分类定位。试验结果表明,对不同样式快递面单均有较好的识别效果,对不良一维码识别率和准确率整体达到60%以上,不同区域三段码定位准确率达到98.47%,单票识别率达到98.03%,整体识别时间在105~146 ms。研究结果可为快递高效分拣提供技术参考。

In order to effectively solve the problems of high labor input cost of the sorting link,slow background information update speed and low sorting efficiency,a simultaneous positioning and recognition method of one-dimensional code and three-segment code based on machine vision was proposed.It mainly includes improving one-dimensional code positioning method and using“line scanning”to improve the decoding accuracy of one-dimensional code.Faster R-CNN target detection method was improved to achieve classification and positioning of three-segment code in different regions.The experimental results show that this method has a good recognition effect on different styles of waybill.The recognition rate and accuracy rate of bad one-dimensional code is more than 60%,the accuracy rate of three-segment code positioning in different regions reaches 98.47%,the overall recognition rate of single waybill reaches 98.03%,and the overall recognition time is between 105 ms and 146 ms.The research results can provide technical reference for efficient sorting of waybills.