基于YOLOX-αSMV的带钢材料表面缺陷检测算法
2024-06-17曹义亲刘文才徐露
曹义亲 刘文才 徐露



摘要:【目的】针对YOLOX算法在钢材表面缺陷检测中特征提取不充分、多目标缺陷检测能力较弱等问题,提出改进损失函数的多维度特征融合带钢材料表面缺陷检测算法。【方法】首先,在Backbone部分应用SPP_SF保留多尺度特征信息,提高分类精度。其次,在Neck部分加入多维度特征融合模块MDFFM,将通道、空间、位置信息融入特征向量中,加强算法的特征提取能力。最后,引入Varifocal Loss和α-CIoU加权正负样本,提高预测框的回归精度。【结果】实验结果表明,YOLOX-αSMV在NEU-DET数据集中的mAP@0.5:0.95达到了47.54%,较YOLOX算法提高了3.43%。【结论】算法在保持检测速度基本不变的情况下,对模糊缺陷和小目标缺陷的识别、定位能力明显提升。
关键词:YOLOX;缺陷检测;α-CIoU;坐标注意力;Varifocal Loss;SoftPool
中图分类号:TP391;U226 文献标志码:A
文章编号:1005-0523(2024)02-0109-09
YOLOX-αSMV Algorithm for Surface Defect Detection
of Strip Steel Material
Cao Yiqin1, Liu Wencai1, Xu Lu2
(1. School of Software, East China Jiaotong University, Nanchang 330013, China; 2. School of Mechanical and Electrical Engineering, Jiangxi Vocational &Technical College of Communications, Nanchang 330013, China)
Abstract: 【Objective】In order to solve the problems of insufficient feature extraction and weak ability of multi-target defect detection of YOLOX algorithm in steel surface defect detection, a multi-dimensional feature fusion strip material surface defect detection algorithm based on improved loss function is proposed. 【Method】First of all, apply SPP_SF to the Backbone part to retain multi-scale feature information and improve classification accuracy. Secondly, the multi-dimensional feature fusion module MDFFM is added in the Neck part to integrate the channel, space and position information into the feature vector to strengthen the feature ex-traction ability of the algorithm. Finally, the introduction of Varifocal Loss and α-CIoU is weighted with positive and negative samples to improve the regression accuracy of the prediction box. 【Result】The experimental results show that YOLOX-αSMV in NEU-DET data set mAP@0.5:0.95 reaches 47.54%, which is 3.43% higher than YOLOX algorithm. 【Conclusion】The algorithm significantly improves the recognition and localization of fuzzy defects and small target defects while keeping the detection speed basically unchanged.
Key words: YOLOX; defect detection; α-CIoU; coordinate attention; Varifocal Loss; SoftPool
Citation format: CAO Y Q, LIU W C, XU L. YOLOX-αSMV algorithm for surface defect detection of strip steel material[J]. Journal of East China Jiaotong University, 2024, 41(2): 109-117.
【研究意义】受生产工艺、环境、使用年限等因素的影响,带钢材料表面会出现表面点蚀、划痕、裂纹、轧制氧化皮等缺陷,导致钢材质量的下降,对工业生产及产品的可靠性和安全性将造成严重影响。因此,如何对带钢材料产品表面的缺陷进行及时有效的检测,一直是相关企业和学者的重要研究内容。
【研究进展】常用的带钢材料表面缺陷检测方法主要有CCD摄像扫描器检测、激光扫描检测、漏磁检测、人工检测等,虽然可以对带钢表面的缺陷进行检测,但存在难度大、成本高、易受外部干扰、精度低、耗时长等一系列问题,很难符合实际检测需求。
机器视觉检测技术凭借成本低、速度快、精度高等优势逐渐成为许多学者的研究热点,并取得了良好效果。……
