| 18 | 0 | 52 |
| 下载次数 | 被引频次 | 阅读次数 |
为了对生产轮胎的缺陷进行精准的检测，提出一种基于改进GANomaly网络的轮胎缺陷检测方法—AAEGAN。首先，通过将编码器替换为对抗自编码器(AAE)，通过对抗训练机制增加对潜在空间的先验分布约束；其次，改变传统的训练策略引入逐步训练策略，显著提升了模型的收敛速度；最后加入感知损失函数，显著提升了模型在无监督条件下的异常检测性能。利用实验室自制的轮胎数据集进行了对比和消融实验，实验结果表明，AUC值提高了6.3%,AP值提高了5.1%，验证了该方法在真实轮胎X光图像数据集上表现出优异的缺陷识别能力。
Abstract：To achieve precise detection of defects in manufactured tires, this study proposes a tire defect detection method based on an improved GANomaly network—AAEGAN. First, by replacing the traditional training strategy with a progressive training strategy, the model′s convergence speed is significantly improved; second, the encoder is replaced with an adversarial autoencoder(AAE), which, through adversarial training, adds a prior distribution constraint to the latent space; finally, a perceptual loss function is introduced, markedly enhancing the model′s anomaly detection performance under unsupervised conditions. Comparative and ablation experiments were conducted using a laboratory-made tire dataset. Experimental results show that the AUC increased by 6.3% and the AP increased by 5.1%, validating the method′s superior defect recognition capability on a real tire X-ray image dataset.
[1] Wang Q, Wu B, Zhu P, et al. ECA-Net:Efficient channel attention for deep convolutional neural networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 44(5):2699-2712.
[2]刘金海,赵真,付明芮,等.基于主动小样本学习的管道焊缝缺陷检测方法[J].仪器仪表学报,2022,43(11):252-261.
[3]罗仁泽,唐祥,余泓,等.基于改进Faster RCNN的射线图像焊缝缺陷检测方法[J].电子测量技术,2023,46(22):160-168.
[4]李大华,徐傲,王笋,等.基于改进YOLOv5的印刷电路板缺陷检测[J].电子测量技术,2023,46(23):112-119.
[5]陈利琼,梅后金,胡洪宣,等.基于改进Faster R-CNN的焊缝缺陷检测方法[J].科学技术与工程,2025,25(5):2027-2033.
[6]陈帅,李焕锋,沙杰,等.基于YOLOv5的砂纸表面缺陷检测方法研究[J].电子测量技术,2023,46(14):73-79.
[7]凌鸿伟,张建敏.基于改进生成对抗网络的晶圆表面缺陷检测[J].现代信息科技,2024,8(20):37-42+47.
[8]孙贺,刘胜波,冷于浩,等.基于机器视觉的轮胎表面缺陷检测系统的研究与应用[J].工业控制计算机,2022,35(01):29-30+34.
[9]刘慧,卢云志,张雷.基于Dropout改进的SRGAN网络DrSRGAN[J].科学技术与工程,2023,23(23):10015-10022.
[10]王振强,杜昊晨,彭泽.改进GANomaly模型的印刷辊筒表面缺陷检测方法[J].信息与电脑(理论版),2022,34(23):190-193+256.
[11] GAO S,DAI Y,XU Y C,et al.Generative adversarial network–assisted image classification for imbalanced tire X-ray defect detection[J].Transactions of the Institute of Measurement and Control,2023,45(8):1492-1504.
[12]宋航,周凤,熊伟.基于自相关-变分对抗学习的物理系统异常检测[J].计算机工程,2024,50(12):358-366.
[13]刘韵婷,刘鑫,高宇.基于FAMGAN的轮胎X光图像缺陷检测[J].电子测量与仪器学报, 2023,37(12):58-66.
[14] Akcay S, Atapour-Abarghouei A, Breckon T P. Skip-GANomaly:Skip connected adversarial autoencoders for anomaly detection[J]. IEEE Transactions on Image Processing, 2019, 28(11):5427-5437.
[15] Schleg L J, et al. Unsupervised anomaly detection with GANs to guide marker discovery[C].International Conference on Information Processing in Medical Imaging(IPMI). Cham:Springer, 2017:146–157.
[16] Akcay S, Atapour-Abarghouei A, Breckon T P. GANomaly:Semi-supervised anomaly detection via adversarial training[C]//Asian Conference on Computer Vision(ACCV). Perth, Australia:Springer, 2018:622-637.
[17] AKCAY S, ATAPOUR-ABARGHOUEI A, BRECKON T P. SkipGANomaly:Skip connected and adversarially trained encoder decoder anomaly detection[C]. Proceedings of the International Joint Conference on Neural Networks. New York:IEEE, 2019:1-8.
[18]刘韵婷,谭明晓,高宇,等.基于双判别器的GANomaly异常检测方法研究[J].电子科技大学学报,2024,53(04):568-575.
[19] Zhang W, Ran L, Deng W, et al. Robust Zero-watermarking for medical image security using residual-based GAN[J]. Journal of King Saud University Computer and Information Sciences,2025,38(2):49.
[20]张果,夏小东,付茂栗,等.基于FastFlow的复杂纹理图像异常检测[J].计算机应用,2024,44(S1):256-261.
[21]税建达,陈龙,卞佰成,等.基于改进PatchCore的无监督异常检测算法[J].上海理工大学学报,2025,47(5):533-541.
基本信息:
中图分类号:U463.341;TP391.41
引用信息:
[1]刘韵婷,李思维,冯欣悦,等.基于逐步训练策略的轮胎X光图像异常检测[J].通信与信息技术,2026,No.283(05):7-13.
2025-10-16
2025
2025-11-25
2025-11-07
2025
1
2026-09-23
2026-09-23