Research on composite material inclusions defect detection based on improved YOLOv8

doi:10.19936/j.cnki.2096-8000.20250628.004

Abstract

Abstract: In order to solve the problem of insufficient detection accuracy caused by small target size and similar features in the detection of composite inclusion defects, an improved YOLOv8 algorithm was proposed. First, the SPD convolution module is introduced to reduce information loss, and the MCA attention mechanism is incorporated to enhance the three-dimensional channel feature extraction capability, thereby improving defect recognition accuracy. Subsequently, the BiFPN bidirectional pyramid network was used to improve the multi-scale feature fusion to improve the model’s ability to identify similar features and size difference defects. Finally, to tackle the bottleneck of small target detection, a Shape IoU loss function is added to optimize the shape and scale of bounding boxes, improving the detection performance for small-size defects. Experimental results show that the improved algorithm achieves a 10.1% increase in mAP@0.5 and a 7.4% increase in mAP@0.95, with an 8.1% improvement in recall rate. The test results in the composite material defect detection system validate the reliability and practicality of this method, providing an efficient and accurate technical solution for composite material inclusion defect detection.

Key words: defect detection, composite materials, YOLOv8

CLC Number:

TB332

WU Zhicheng, WANG Mingquan, XIE Shaopeng, LU Yupeng, CAO Zhenfeng, WANG Jinhua. Research on composite material inclusions defect detection based on improved YOLOv8[J]. COMPOSITES SCIENCE AND ENGINEERING, 2025, 0(6): 27-33.

References

[1] BOCHKOVSKIY A, WANG C-Y, LIAO H-Y M. YOLOv4: optimal speed and accuracy of object detection[EB/OL]. (2020-04-23)[2024-12-30]. https://arxiv.org/pdf/2004.10934v1.
[2] LI C Y, LI L L, JIANG H L, et al. YOLOv6: a single-stage object detection framework for industrial applications[EB/OL]. (2022-09-07)[2024-09-14]. https://arxiv.org/abs/2209.02976.
[3] REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]//IEEE. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas: IEEE, 2016: 779-788.
[4] REDMON J, FARHADI A. YOLO9000: better, faster, stronger[C]//IEEE. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI, USA: IEEE, 2017: 6517-6525.
[5] REDMON J, FARHADI A. YOLOv3: an incremental improvement[EB/OL]. (2018-04-08)[2024-12-30]. https://arxiv.org/abs/1804.02767.
[6] 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[C]//IEEE. 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Vancouver, BC, Canada: IEEE, 2023: 7464-7475.
[7] 苏志威, 黄子涵, 邱发生, 等. 基于改进YOLOv8的航空铝合金焊缝缺陷检测方法[J]. 航空动力学报, 2024, 39(6): 121-129.
[8] 谢雨欣, 龚烨飞, 谷心浩, 等. 基于RGB-D特征融合的焊缝表面缺陷检测方法[J]. 焊接学报, 2024, 45(12): 72-78.
[9] SUNKARA R, LUO T. No more strided convolutions or pooling: a new CNN building block for low-resolution images and small objects[C]//Joint European Conference on Machine Learning and Knowledge Discovery in Databases. Cham: Springer Nature Switzerland, 2022: 443-459.
[10] YU Y, ZHANG Y, et al. MCA: multidimensional collaborative attention in deep convolutional neural networks for image recognition[J]. Engineering Applications of Artificial Intelligence, 2023, 126:107079.
[11] TAN M X, PANG R M, LE Q V. Efficient Det: scalable and efficient object detection[C]//IEEE. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle: IEEE, 2020: 10781-10790.
[12] ZHANG H, ZHANG S. Shape-IoU: more accurate metric considering bounding box shape and scale[EB/OL]. (2023-12-29)[2024-12-30]. https://arxiv.org/abs/2312.17663.
[13] MA M, PANG H. SP-YOLOv8s: an improved YOLOv8s model for remote sensing image tiny object detection[J]. Applied Sciences, 2023, 13(14): 8161.
[14] 司坤宇, 牛春晖. 基于混合差分卷积和高效视觉Transformer网络的三重多模态图像融合算法[J]. 红外与激光工程, 2024, 53(11): 331-345.
[15] LIN T-Y, DOLLÁR P, GIRSHICK R, et al. Feature pyramid networks for object detection[C]//IEEE. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI, USA: IEEE, 2017: 936-944.
[16] LIU S, QI L, QIN H, et al. Path aggregation network for instance segmentation[C]//IEEE. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA: IEEE, 2018: 8759-8768.
[17] GHIASI G, LIN T Y, LE Q V.NAS-FPN: learning scalable feature pyramid architecture for object detection[C]//IEEE. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Long Beach: IEEE, 2019: 7029-7038.
[18] DU S, ZHANG B, ZHANG P, et al. An improved bounding box regression loss function based on CIOU loss for multi-scale object detection[C]//IEEE. 2021 IEEE 2nd International Conference on Pattern Recognition and Machine Learning (PRML). Chengdu: IEEE, 2021: 92-98.
[19] 李磊磊, 王明泉, 赵付宝, 等. 基于Mask R-CNN的复合材料夹杂缺陷自动检测研究[J]. 复合材料科学与工程, 2024(1): 83-88.
[20] REN S, HE K, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2017, 39(6): 1137-1149.
[21] 彭鸿瑞, 杨桂华. 基于改进YOLOv8的SOP芯片缺陷检测研究[J]. 电子测量技术, 2024, 47(12): 71-82.