改进Yolov7-tiny的钢材表面缺陷检测算法

doi:10.3778/j.issn.1002-8331.2302-0191

摘要/Abstract

摘要： 为提高钢材表面缺陷小目标检测效率，提出一种改进Yolov7-tiny的钢材表面缺陷检测算法。将特征提取网络的激活函数更换为SiLU，提高特征提取能力；把特征融合网络的张量拼接操作与加权双向特征金字塔BiFPN结合，再把上采样部分的最邻近插值替换为轻量级算子CARAFE，提升特征融合能力；最后在输出端引入多头自注意力机制MHSA 和SPD卷积构建块，提升输出端对钢材表面缺陷小目标的检测性能。在NEU-DET数据集上做消融和对比实验，改进算法与原Yolov7-tiny算法比较，mAP提升11.7个百分点，Precision提升3.3个百分点，FPS值达到192，结果表明改进算法能有效提升钢材表面缺陷小目标检测效率；在VOC2012数据集上做通用性对比实验，结果表明改进算法具有通用性。

关键词: 钢材表面, 缺陷检测, Yolov7-tiny, SiLU, BiFPN, CARAFE, MHSA, SPD

Abstract: In order to improve the efficiency of small target detection of steel surface defects, an improved Yolov7-tiny steel surface defect detection algorithm is proposed. The activation function of the feature extraction network is changed to SiLU to improve the feature extraction capability. The tensor splicing operation of the feature fusion network is combined with the weighted bidirectional feature pyramid BiFPN, and the nearest interpolation of the upper sampling part is replaced with the lightweight operator CARAFE to improve the feature fusion ability. Finally, the multi-head self-attention mechanism MHSA and SPD convolution building blocks are introduced at the output end to improve the detection performance of the output end for small targets of steel surface defects. The ablation and contrast experiments are carried out on the NEU-DET dataset. Compared with the original Yolov7-tiny algorithm, the improved algorithm has increased the mAP by 11.7 percentage points, the precision by 3.3 percentage points, and the FPS value reaches 192. The results show that the improved algorithm can effectively improve the detection efficiency of small targets of steel surface defects. Comparative experiments on the VOC2012 dataset show that the improved algorithm is universal.

Key words: steel surface, defect detection, Yolov7-tiny, SiLU, BiFPN, CARAFE, MHSA, SPD

齐向明, 董旭. 改进Yolov7-tiny的钢材表面缺陷检测算法[J]. 计算机工程与应用, 2023, 59(12): 176-183.

QI Xiangming, DONG Xu. Improved Yolov7-tiny Algorithm for Steel Surface Defect Detection[J]. Computer Engineering and Applications, 2023, 59(12): 176-183.

参考文献

[1] WANG Y，LIAN X F，LU Y.Product defects detection based on features fusion and the multi-scale sliding window[C]//2017 13th International Conference on Natural Computation，Fuzzy Systems and Knowledge Discovery（ICNC-FSKD），2017：2948-2952.
[2] CHU M X，LIU X P，GONG R F，et al.Multi-class classification method using twin support vector machines with multi-information for steel surface defects[J].Chemometrics and Intelligent Laboratory Systems，2018，176：108-118.
[3] 刘坤，刘铁旭，陈海永，等.基于自适应排序融合注意力网络的带钢表面缺陷检测[C]//中国自动化大会论文集，2021：724-729.
LIU K，LIU T X，CHEN H Y，et al.Strip steel surface defect detection based on adaptive sort fusion attention network[C]//Proceedings of China Automation Conference，2021：724-729.
[4] 方钧婷，谭晓阳.注意力级联网络的金属表面缺陷检测算法[J].计算机科学与探索，2021，15（7）：1245-1254.
FANG J T，TAN X Y.Defect detection of metal surface based on attention cascade R-CNN[J].Journal of Frontiers of Computer Science and Technology，2021，15（7）：1245-1254.
[5] 马宇超，付华良，吴鹏，等.深度网络自适应优化的Mask R-CNN模型在铸件表面缺陷检测中的应用研究[J].现代制造工程，2022（4）：112-118.
MA Y C，FU H L，WU P，et al.Research on the Mask R-CNN model of deep network adaptive optimization in the detection of casting surface defects[J].Modern Manufacturing Engineering，2022（4）：112-118.
[6] 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，2016：779-788.
[7] REDMON J，FARHADI A.Yolov3：an incremental improvement[J].arXiv：1804.02767，2018.
[8] BOCHKOVSKIY A，WANG C Y，LIAO H Y M.Yolov4：optimal speed and accuracy of object detection[J].arXiv：2004.10934，2020.
[9] 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.
[10] LIAN J，HE J，NIU Y，et al.Fast and accurate detection of surface defect based on improved YOLOv4[J].Assembly Automation，2021.
[11] GE Z，LIU S，WANG F，et al.Yolox：exceeding yolo series in 2021[J].arXiv：2107.08430，2021.
[12] LI C，LI L，JIANG H，et al.YOLOv6：a single-stage object detection framework for industrial applications[J].arXiv：2209.02976，2022.
[13] WANG C Y，BOCHKOVSKIY A，LIAO H Y M.YOLOv7：trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[J].arXiv：2207.02696，2022.
[14] LIN T Y，GIRSHICK R.Feature pyramid networks for object detection[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition.New York：IEEE Press，2017：936-944.
[15] LIU S，QI L，QIN H F，et al.Path aggregation network for instance segmentation[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.New York：IEEE Press，2018：8759-8768.
[16] WANG C Y，YEH I H，LIAO H Y M.You only learn one representation：unified network for multiple tasks[J].arXiv：2105.04206，2021.
[17] TAN M X，PANG R M，LE Q V.EfficientDet：scalable and efficient object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2020：10781-10790.
[18] WANG J Q，CHEN K，XU R，et al.Carafe：content-aware reassembly of features[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision，2019：3007-3016.
[19] SRINIVAS A，LIN T Y，PARMAR N，et al.Bottleneck transformers for visual recognition[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition，2021：16519-16529.
[20] SUNKARA R，LUO T.No more strided convolutions or pooling：a new CNN building block for low-resolution images and small objects[J].arXiv：2208.03641，2022.
[21] SAJJADI M S M，VEMULAPALLI R，BROWN M.Frame-recurrent video super-resolution[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，2018：6626-6634.