With the deep integration of new generation information technology and the manufacturing industry,it has triggered tremendous changes in the manufacturing industry, and the detection of surface defects in industrial production has become increasingly important. A YOLOv8-based improved model algorithm is proposed to address the issues of missed and false detections, as well as inaccurate detection accuracy in steel surface defect detection. Firstly, in the backbone network, the original C2f module has been integrated with the spatial and channel reconstruction convolution module to reduce inherent redundancy in dense model parameters and further develop lightweight network models. Secondly, the path aggregation network is replaced by the bidirectional feature pyramid network in the Neck stage, which is intended to improve the detection performance of algorithms by using effective feature fusion methods. Finally, CIOU in YOLOv8 is replaced by Wise-IOUv1 to optimize the loss function. Experimental results show that the improved YOLOv8 algorithm has increased mAP by 8.1 percentage points, reduced computational complexity by 9.6 percentage points, and improved detection speed by 8.7 percentage points, compared to the original algorithm. This detection performance is better than the original algorithm, which greatly improved the detection performance of steel surface defect detection.
GAOChunyan, QINShen, LIManhong, et al. Research on steel surface defect detection with improved YOLOv7 algorithm[J]. Computer Engineering and Applications, 2024, 60(7): 282-291. (in Chinese)
GONGChenbo, ZhuojiangNAN, TAOWei. Research on PCB surface defect detection based on improved Faster RCNN[J]. Process Automation Instrumentation, 2024, 45(7): 99-103. (in Chinese)
XUXinyu, SHENTong, JiaLÜ. Steel surface defect detection based on improved YOLOv8 algorithm[J]. Automation Application, 2024(15): 6-10. (in Chinese)
[16]
DUANC, WANGX, JIL, et al. Research on defect detection method for end face of steel coil based on Faster-RCNN and YOLOv3[C]//International Conference on Neural Networks, Information, and Communication Engineering (NNICE 2022), 2022: 1-8.
[17]
DINGF, ZHUANGZ, LIUY, et al. Detecting defects on solid wood panels based on an improved SSD algorithm[J]. Sensors, 2020, 20(18): 5315.
YAOJingli, CHENGGuang, WANFei, et al. Improved lightweight bearing defect detection algorithm of YOLOv8[J]. Computer Engineering and Applications, 2024, 60(21): 205-214. (in Chinese)
[23]
FENGW, LIUM, SUNY, et al. The use of a blueberry ripeness detection model in dense occlusion scenarios based on the improved YOLOv9[J]. Agronomy, 2024, 14(8): 1860.
[24]
TIE J, ZHUC, ZHENGL, et al. LSKA-YOLOv8: a lightweight steel surface defect detection algorithm based on YOLOv8 improvement[J]. Alexandria Engineering Journal, 2024, 109: 201-212.
LUOLiang, LANGXiao, ZUGuoqing, et al. A Cylinder liner defect detection method based on improved YOLOv8[J/OL]. China Mechanical Engineering, 2024: 1-12. in Chinese)
[29]
YEY, LIUQ, LIL, et al. Improving insulator fault detection with effective-YOLOv7 network[J]. Journal of Electronic Imaging, 2023, 32(6): 063021.