基于YOLOv4网络模型的金属表面划痕检测

doi:10.12382/bgxb.2022.A011

摘要/Abstract

摘要： 金属表面划痕检测旨在实现金属划痕的分类识别和尺度精确定位。由于划痕本身形态各异且其与背景的低对比度特性，当前基于传统图像处理方法难以精确实现划痕的定位与识别。为此提出一种基于小尺度卷积核的浅层神经网络模型，模型借鉴小目标检测相关理论，在数据层面对划痕进行增强，训练网络模型，实现表面划痕的精确检测。结果表明，相对于原始的YOLOv4网络模型而言，该网络模型且能够更好地避免不明显划痕的漏检测和误检测现象的出现，也能够更精确且完整地提取出贯穿式或较长的划痕。该模型完全能够满足生产线精确检验要求。

关键词: 金属表面, 划痕检测, YOLOv4网络模型, NEU数据集

Abstract: Metal surface scratch detection aims to achieve the classification recognition and precise scale localization of metal scratchs. The current traditional image processing methods are difficult to achieve the accurate localization and recognition of scratches due to their different shapes and low contrast with the background.A shallow neural network model based on small scale convolution kernel is proposed. The proposed model uses the correlation theory of small target detection to enhance the scratches at the data level first and then train the network model to achieve the accurate detection of surface scratches.Compared with the original YOLOv4 model， the proposed model can avoid the missing detection and false detection of insignificant scratches better， and also can extract the penetrating or long scratches more accurately and completely.The model can fully meet the requirements of accurate inspection of production line.

Key words: metalsurface, scratchdetection, YOLOv4networkmodel, NEUdataset

中图分类号:

TP391.4

张博尧，冷雁冰. 基于YOLOv4网络模型的金属表面划痕检测[J]. 兵工学报, 2022, 43(S1): 214-221.

ZHANG Boyao， LENG Yanbing. Detection of Metal Surface Scratch Based on YOLOv4 Network Model[J]. Acta Armamentarii, 2022, 43(S1): 214-221.

参考文献

［1］ GHORAI S，MUKHERJEE A，GANGADARAN M，et al.Automatic defect detection on hot-rolled flat steel products［J］.

IEEE Transactions on Instrumentation and Measurement，2012，62(3):612-621.
［2］ SONG K C，YAN Y H.A noise robust method based on completed local binary patterns for hot-rolled steel strip surface defects［J］.Applied Surface Science，2013，285(Part B):858-864.
［3］ ZHANG H，JIN X T，WU Q M J，et al.Automatic visual detection system of railway surface defects with curvature filter and improved Gaussian mixture model［J］.IEEE Transactions on Instrumentation and Measurement，2018，67(7):1593-1608.
［4］ WANG H Y，ZHANG J W，TIAN Y，et al.A simple guidance template-based defect detection method for strip steel surfaces［J］.IEEE Transactions on Industrial Informatics，2018，15(5):2798-2809.
［5］ GIRSHICK R.Fast R-CNN［C］∥Proceedings of the IEEE International Conference on Computer Vision.Santiago，Chile:IEEE，2015:1440-1448.
［6］ HE K M，GKIOXARI G，DOLLR P，et al.Mask R-CNN［C］∥Proceedings of the IEEE International Conference on Computer Vision. Venice，Italy：IEEE，2017:2980-2988.
［7］ REN S，HE K，GIRSHICK R，et al.Faster R-CNN:towards real-time object detection with region proposal networks［J］.arXiv preprint arXiv，2015:1506.01497.

表5 金属表面划痕检测结果视觉比较（优化模型提取的目标范围信息较为完整，
原始模型只提取到划痕的一部分）
Tab.5 Visual comparison of metal surface scratch detection results (the target range information extracted from
the optimized model is relatively complete， and the original model only extracted a part of the scratches

［8］ REDMON J，DIVVALA S，GIRSHICK R，et al.You only look once: unified， real-time object detection［C］∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Las Vegas，NV，US: IEEE， 2016.
［9］ REDMON J，FARHADI A.YOLO 9000:better，faster，stronger［C］∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Honolulu，HI，US:IEEE，2017.
［10］ REDMON J，FARHADI A.YOLOv3:An incremental improvement［J］.arXiv preprint arXiv，2018：02767.
［11］ BOCHKOVSKIY A，WANG C Y，LIAO H Y M. YOLOv4: Optimal speed and accuracy of object detection［J］.arXiv preprint arXiv，2020：10934.
［12］张永梅，赖裕平，马健喆，等.基于视频的装甲车和飞机检测跟踪及轨迹预测算法［J］.兵工学报，2021，42（3）：545- 554.
ZHANG Y M，LAI Y P，MA J Z，et al. A video-based prediction algorithm for armored vehicle and aircraft detection/tracking and trajectory［J］.Acta Armamentarii，2021，42（3）：545-554.(in Chinese)
［13］穆思奇，林进健，汪海泉，等. 基于改进YOLOv4的X射线图像违禁品检测算法［J］. 兵工学报，2021，42（12）：2675-2683.
MU S Q，LIN J J，WANG H Q，et al. An algorithm for detection of prohibited items in X-ray images based on improved YOLOv4［J］.Acta Armamentarii，2021，42（12）：2675-2683.(in Chinese)
［14］ BOCHKOVSKIY A，WANG C Y，LIAO H Y M.YOLOv4:Optimal speed and accuracy of object detection［J］.arXiv preprint arXiv，2020：10934.
［15］ MASCI J，MEIER U，FRICOUT G，et al.Multi-scale pyramidal pooling network for generic steel defect classification［C］∥Proceedings of the 2013 International Joint Conference on Neural Networks. Dallas，TX，US:IEEE，2013.
［16］ NATARAJAN V，HUNG T Y，VAIKUNDAM S，et al.Convolutional networks for voting-based anomaly classification in metal surface inspection［C］∥Proceedings of the 2017 IEEE International Conference on Industrial Technology.Toronto，ON，Canada:IEEE，2017:986-991.
［17］ HE Y，SONG K C，DONG H W，et al.Semi-supervised defect classification of steel surface based on multi-training and generative adversarial network［J］.Optics and Lasers in Engineering，2019，122:294-302.
［18］ HE Y，SONG K C，MENG Q G，et al.An end-to-end steel surface defect detection approach via fusing multiple hierarchical features［J］.IEEE Transactions on Instrumentation and Measurement，2019，69(4):1493-1504.
［19］ JOSESH N， JACOB S. Responding to the controversy about YOLOv5［EB/OL］. (2022-06-13)［2022-06-13］. https:∥blog.roboflow.com/yolov4-versus-yolov5.
［20］ HE K M，ZHANG X Y，REN S Q，et al.Spatial pyramid pooling in deep convolutional networks for visual recognition［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence，2015，37(9): 1904-1916.
［21］ LIU S，QI L，QIN H F，et al.Path aggregation network for instance segmentation［C］∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Salt Lake City，UT，US:IEEE， 2018:8759-8768.
［22］ WANG C Y，LIAO H Y M，WU Y H，et al.CSPNet: A new backbone that can enhance learning capability of CNN［C］∥Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops.Seattle，WA，US:IEEE，2020.

[1]	李博，王博，韩京冶，杨宗睿，栗霖雲，傅畅之. 基于车载计算机的红外图像移动目标检测[J]. 兵工学报, 2022, 43(S1): 66-73.
[2]	张良安, 陈洋, 谢胜龙, 刘同鑫. 基于机器视觉与深度学习的飞机防护栅裂纹检测系统[J]. 兵工学报, 0, (): 0-0.
[3]	霍健, 陈慧敏, 马云飞, 郭鹏宇, 杨旭, 孟祥盛. 基于MEMS激光雷达的车辆目标识别算法[J]. 兵工学报, 0, (): 0-0.
[4]	钱立志, 杨传栋, 薛松, 陈栋, 凌冲. 图像自寻的弹药目标检测方法综述[J]. 兵工学报, 0, (): 0-0.
[5]	唐伟, 贾方秀, 王晓鸣. 基于双边滤波的可见光与红外图像自适应融合[J]. 兵工学报, 0, (): 0-0.
[6]	于博文，吕明. 改进的YOLOv3算法及其在军事目标检测中的应用[J]. 兵工学报, 2022, 43(2): 345-354.
[7]	张凤, 张超, 杨华民, 王发斌. 作战辅助系统中基于深度学习的投影图像颜色补偿方法[J]. 兵工学报, 2021, 42(11): 2418-2423.
[8]	才华，孙俊，朱瑞昆，朱新丽，赵义武. 基于自适应圆边际的深度人脸识别算法[J]. 兵工学报, 2021, 42(11): 2424-2432.
[9]	丁贵鹏，陶钢，李英超, 庞春桥，王小峰，段桂茹. 基于非下采样轮廓波变换与引导滤波器的红外及可见光图像融合[J]. 兵工学报, 2021, 42(9): 1911-1922.
[10]	姜珊，底晓强，韩成. 融合时空特性的孪生网络视觉跟踪[J]. 兵工学报, 2021, 42(9): 1940-1950.
[11]	张永梅，赖裕平，马健喆，冯超，束颉. 基于视频的装甲车和飞机检测跟踪及轨迹预测算法[J]. 兵工学报, 2021, 42(3): 545-554.
[12]	景文博，黄炳坤，刘健，于洪洋. 激光照射性能测量中异源相机的标定方法[J]. 兵工学报, 2020, 41(5): 924-931.
[13]	吴成茂，孙佳美. 基于全散度的自适应鲁棒图形模糊聚类算法[J]. 兵工学报, 2019, 40(9): 1890-1901.
[14]	李春明，姜雨彤，宋海平，纪超，郭猛，朱琳. 基于光学厚度代理模型的雾浓度估计及图像去雾[J]. 兵工学报, 2019, 40(7): 1425-1433.
[15]	李从利，张思雨，韦哲，薛松. 基于深度卷积生成对抗网络的航拍图像去厚云方法[J]. 兵工学报, 2019, 40(7): 1434-1442.