本文二维码信息
二维码（扫一下试试看！）
基于YOLO v3的卷烟用瓦楞纸箱表面缺陷检测方法
Surface Defect Detection Method for Corrugated Box for Cigarettes Based on YOLO v3 Algorithm
收稿日期：2022-11-22  
DOI：10.11980/j.issn.0254-508X.2023.06.017
关键词:  卷烟用瓦楞纸箱回收  缺陷检测  YOLO v3
Key Words:cigarette corrugated box recycling  defect detection  YOLO v3
基金项目:
作者单位邮编
贾伟萍 湖北中烟工业有限责任公司，湖北武汉，430000 430000
褚玮 湖北中烟工业有限责任公司，湖北武汉，430000 430000
刘文婷 湖北中烟工业有限责任公司，湖北武汉，430000 430000
黄轲 湖北中烟工业有限责任公司，湖北武汉，430000 430000
李陈巧 湖北中烟工业有限责任公司，湖北武汉，430000 430000
吴飞* 武汉理工大学，湖北武汉，430070 430070
摘要点击次数: 981
全文下载次数: 766
摘要:提出了一种基于YOLO v3的检测算法，并建立了卷烟用瓦楞纸箱循环利用性能评价体系。对比了Faster RCNN与YOLO v3深度神经网络目标检测算法对采集的典型卷烟用瓦楞纸箱表面缺陷的识别结果。基于OpenCV库、Canny算法，开发了适用于测量卷烟用瓦楞纸箱表面缺陷尺寸及其分布位置的合理方案。结果表明，该算法的平均准确率达92.23%，可成功实现缺陷位置、尺寸的检测。
Abstract:A detection algorithm based on YOLO v3 was proposed， and an evaluation system for the recycling performance of corrugated box for cigarettes was established. The recognition results of surface defects of collected typical corrugated box for cigarettes using Faster RCNN and YOLO v3 deep neural network target detection algorithms were compared. Based on the OpenCV library and Canny algorithm， a reasonable solution suitable for measuring the sizes and distribution locations of surface defects of corrugated box for cigarettes was developed. The results showed that the algorithm had an average accuracy of 92.23% and could successfully achieve the detection of defect locations and size.
查看全文  HTML  查看/发表评论  下载PDF阅读器