改进FCOS模型的微通道铝扁管表面缺陷检测算法

doi:10.3778/j.issn.1002-8331.2208-0204

摘要/Abstract

摘要： 针对微通道铝扁管表面缺陷的检测问题，提出了一种改进的FCOS（fully convolutional one-stage object detection）算法。设计一种特征卷积金字塔网络（feature convolutional pyramid network，FCPN），使模型能够自适应地混合不同层的特征图进行检测；在分析原始FCOS应用于狭长缺陷检测局限性的基础上，改进了模型的正样本部署策略，以降低模型对狭长缺陷的漏检；设计更加合适的映射函数与中心度函数，解决标注框外正样本点的回归问题与中心度计算问题；使用EIoU损失（efficient IoU loss）替换原模型中的IoU损失，进一步提高模型的回归能力。实验结果表明，在微通道铝扁管的表面缺陷检测任务中，改进后的FCOS模型达到了76.4%的mAP（mean average precision），相比于原始模型提高了7.7个百分点。

关键词: 微通道铝扁管, 狭长缺陷检测, 样本部署, 特征卷积金字塔, FCOS模型

Abstract: Aiming at the task of detecting surface defects of micro-channel aluminum flat tubes, an improved FCOS（fully convolutional one-stage object detection） algorithm is proposed. Firstly, a feature convolution pyramid network is designed to enable the model to adaptively mix feature maps from different layers for detection. Secondly, through analyzing the limitations of the original FCOS algorithm in the detection of long and narrow defects, and the positive sample deployment strategy of the model is improved, the missed detection of long and narrow defects is reduced. Then, a more suitable mapping function and center-ness function are designed to solve the regression problem and center-ness calculation problem of positive sample points outside the labeled frame. Finally, the EIoU（efficient IoU） loss is used to replace the IoU loss in the original model to further improve the regression ability of the model. The experimental results show that in the surface defect detection task of micro-channel aluminum flat tubes, the improved FCOS model achieves the mAP（mean average precision） of 76.4%, which is 7.7 percentage points higher than the original model.

Key words: micro-channel aluminum flat tube, detection of long and narrow defects, sample deployment, feature convolutional pyramid, fully convolutional one-stage object detection（FCOS） model

桂鹏辉, 宋涛, 汤建斌, 徐志鹏, 曹松晓, 蒋庆. 改进FCOS模型的微通道铝扁管表面缺陷检测算法[J]. 计算机工程与应用, 2023, 59(24): 298-308.

GUI Penghui, SONG Tao, TANG Jianbin, XU Zhipeng, CAO Songxiao, JIANG Qing. Surface Defect Detection Algorithm of Micro-Channel Aluminum Flat Tube Based on Improved FCOS Model[J]. Computer Engineering and Applications, 2023, 59(24): 298-308.

参考文献

[1] BOCHKOVSKIY A，WANG C Y，LIAO H Y.Yolov4：optimal speed and accuracy of object detection[J].arXiv：2004.
10934，2020.
[2] 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（CVPR），Las Vegas，NV，USA，2016：779-788.
[3] REDMON J，FARHADI A.YOLO9000：better，faster，stronger[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition（CVPR），Honolulu，HI，USA，2017：7263-7271.
[4] REDMON J，FARHADI A.Yolov3：an incremental improvement[J].arXiv：1804.02767，2018.
[5] LIU W，ANGUELOV D，ERHAN D，et al.SSD：single shot multibox detector[C]//European Conference on Computer Vision（ECCV），Amsterdam，Netherlands，2016：21-37.
[6] REN S，HE K，GIRSHICK R，et al.Faster R-CNN：towards real-time object detection with region proposal networks[J].arXiv：1506.01497，2015.
[7] 程婧怡，段先华，朱伟.改进YOLOv3的金属表面缺陷检测研究[J].计算机工程与应用，2021，57（19）：252-258.
CHENG J Y，DUAN X H，ZHU W.Research on metal surface defect detection by improved YOLOv3[J].Computer Engineering and Applications，2021，57（19）：252-258.
[8] ARTHUR D，VASSILVITSKII S.[k]-means++：the advantages of careful seeding[C]//Proceedings of the Eighteenth Annual ACM-SIAM Symposium on Discrete Algorithms（SODA’07），January 2007：1027-1035.
[9] ZHENG Z，WANG P，LIU W，et al.Distance-IoU loss：faster and better learning for bounding box regression[C]//Proceedings of the AAAI Conference on Artificial Intelligence，New York，USA，2020：12993-13000.
[10] 李维刚，叶欣，赵云涛，等.基于改进YOLOv3算法的带钢表面缺陷检测[J].电子学报，2020，48（7）：1284-1292.
LI W G，YE X，ZHAO Y T，et al.Strip steel surface defect detection based on improved YOLOv3 algorithm[J].Acta Electronica Sinica，2020，48（7）：1284-1292.
[11] 代小红，陈华江，朱超平.一种基于改进Faster RCNN的金属材料工件表面缺陷检测与实现研究[J].表面技术，2020，49（10）：362-371.
DAI X H，CHEN H J，ZHU C P.Surface defect detection and realization of metal workpiece based on improved Faster RCNN[J].Surface Technology，2020，49（10）：362-371.
[12] HE K，GKIOXARI G，DOLLáR P，et al.Mask R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision（ICCV），Venice，Italy，2017：2961-2969.
[13] TIAN Z，SHEN C，CHEN H，et al.Fcos：fully convolutional one-stage object detection[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision（ICCV），Seoul，Korea（South），2019：9627-9636.
[14] ZHOU D，FANG J，SONG X，et al.Iou loss for 2d/3d object detection[C]//2019 International Conference on 3D Vision（3DV），Québec，Canada，2019：85-94.
[15] ZHANG Y F，REN W，ZHANG Z，et al.Focal and efficient IoU loss for accurate bounding box regression[J].arXiv：2101.08158，2021.
[16] HE K，ZHANG X，REN S，et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition（CVPR），Las Vegas，NV，USA，2016：770-778.
[17] LIN T Y，DOLLáR P，GIRSHICK R，et al.Feature pyramid networks for object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition（CVPR），Honolulu，HI，USA，2017：2117-2125.
[18] HUANG Z，WANG X，HUANG L，et al.Ccnet：criss-cross attention for semantic segmentation[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision（ICCV），Seoul，Korea（South），2019：603-612.
[19] WU Y，HE K.Group normalization[C]//Proceedings of the European Conference on Computer Vision（ECCV），Munich，Germany，2018：3-19.
[20] AGARAP A F.Deep learning using rectified linear units（ReLU）[J].arXiv：1803.08375，2018.
[21] TIAN Z，SHEN C，CHEN H，et al.Fcos：a simple and strong anchor-free object detector[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2022，44（4）：1922-1933.
[22] CLEVERT D A，UNTERTHINER T，HOCHREITER S.Fast and accurate deep network learning by exponential linear units（ELUs）[J].arXiv：1511.07289，2015.
[23] MUHAMMAD M B，YEASIN M.Eigen-CAM：class activation map using principal components[C]//2020 International Joint Conference on Neural Networks（IJCNN），2020：1-7.
[24] MAAS A L，HANNUN A Y，NG A Y.Rectifier nonlinearities improve neural network acoustic models[C]//Proceedings of IEEE International Conference on Machine Learning，Atlanta，USA，2013.