2021年 05期

Surface Defect Detection of Industrial Products Based on Faster Region Convolutional Neural Network Model


摘要(Abstract):

针对传统带钢表面缺陷检测算法检测效率低、准确率差的情况,提出一种基于快速选区卷积神经网络模型的多尺度带钢表面缺陷检测算法;首先利用残差网络思想对该模型网络特征提取层进行改进;其次,利用多尺度推荐区域网络设置合理大小的卷积滑动窗口,提取出更加准确的推荐区域;最后,利用软判决非极大值抑制机制替代传统的非极大值抑制机制,解决缺陷特征相近时检测框丢失的情况,并在SD_data数据集上进行实验验证。结果表明,所提出的算法对多尺度带钢表面缺陷的检测准确率明显提高,漏检率显著降低。

关键词(KeyWords): 表面缺陷检测;多尺度推荐区域网络;卷积神经网络

基金项目(Foundation): 国家自然科学基金项目(51705289);; 山东省重点研发计划项目(2019GGX104101);; 山东大学教学研究项目(2019Y112)

作者(Author): 王李馥颖,朱振杰,杜付鑫

DOI: 10.13349/j.cnki.jdxbn.20210601.001

参考文献(References):

[1] 李晓.我国炼钢工艺低碳技术发展方向[J].冶金经济与管理,2019(4):21-24.

[2] 茅正冲,时文静,邬锋.带钢表面缺陷检测方法[J].计算机工程与设计,2014,35(1):233-236,242.

[3] 徐科,宋敏,杨朝霖,等.隐马尔可夫树模型在带钢表面缺陷在线检测中的应用[J].机械工程学报,2013,49(22):34-40.

[4] MAO Q C,SUN H M,LIU Y B,et al.Mini-YOLOv3:real-time object detector for embedded applications[J].IEEE Access,2019,7:133529-133538.

[5] ZUO L Q,SUN H M,MAO Q C,et al.Natural scene text recognition based on encoder-decoder framework[J].IEEE Access,2019,7:62616-62623.

[6] 王宏,朱德生,唐威.一种基于灰度投影的带钢表面缺陷检测算法[J].东北大学学报(自然科学版),2008,29(3):375-377.

[7] 汤勃,孔建益,伍世虔.机器视觉表面缺陷检测综述[J].中国图象图形学报,2017,22(12):1640-1663.

[8] AHMED R,SUTCLIFFE M P F.Identification of surface features on cold-rolled stainless-steel strip[J].Wear,2000,244(1/2):60-70.

[9] GIRSHICK R,DONAHUE J,DARRELL T,et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//2014 IEEE Conference on Computer Vision and Pattern Recognition,June 23-28,2014,Columbus.New York:IEEE,2014:580-587.

[10] ZHANG Q X,LIN G H,ZHANG Y M,et al.Wildland forest fire smoke detection based on faster R-CNN using synthetic smoke images[J].Procedia Engineering,2018,211:441-446.

[11] REN S Q,HE K M,GIRSHICK R,et al.Faster R-CNN:towards real-time object detection with region proposal networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2017,39(6):1137-1149.

[12] LYU Y H,ZHANG C,YUN W J,et al.The delineation and grading of actual crop production units in modern smallholder areas using RS data and mask R-CNN[J].Remote Sensing,2020,12(7):1074-1086.

[13] XUE Y J,HUANG N,TU S Q,et al.Immature mango detection based on improved YOLOv2[J].Transactions of the Chinese Society of Agricultural Engineering,2018,34(7):173-179.

[14] 梁爽,何永辉,杨水山,等.带钢表面缺陷在线检测设备的防护分析与应用[J].世界钢铁,2012,12(2):50-55.

[15] 刘华,顾廷权,缪明华.带钢断面形状检测与特征参数识别方法[J].钢铁研究学报,2012,24(2):59-62.

[16] TAWSIK JAWAD K M,RASHID B M,SAKIB N.Targeted face recognition and alarm generation for security surveillance using single shot multibox detector (SSD)[J].International Journal of Computer Applications,2019,177(22):8-13.

[17] 何欣颖,吴黎明,郑耿哲,等.基于Inception-ResNet-v2的乳腺癌辅助诊断方法[J].自动化与信息工程,2020,41(1):16-21.

[18] YANG X H,ZHAO P P,DONG Y,et al.An improved recombinase polymerase amplification assay for visual detection of Vibrio parahaemolyticus with lateral flow strips[J].Journal of Food Science,2020,85(6):1834-1845.