首页 > 分享 > 基于FACR深度迁移网络的叶轮表面缺陷检测方法

基于FACR深度迁移网络的叶轮表面缺陷检测方法

花匠小妙招
2025-06-27 11:06

随着工业智能化的发展,叶轮表面缺陷检测在保障设备安全与提升生产效率中扮演着关键角色。然而,实际工业环境下缺陷样本信息不完备、样本标记耗时以及检测模型适应性差等问题制约着检测精度的提升。为解决这些问题,提出基于傅里叶约束的自适应一致性正则化（FACR）网络的叶轮表面缺陷检测方法。研究基于傅里叶变换的隐式信息约束策略,针对不同域的缺陷图片进行傅里叶变换得到相位和振幅,对振幅进行线性插值以增强样本信息,同时保证相位不变交换振幅的对应区域,使得模型关注相位信息（即结构信息）,降低背景环境对检测模型的影响;引入课程学习思想动态调整阈值,解决固定阈值无法考虑不同类别的不同学习状态和学习难度导致复杂缺陷检测精度低的问题;结合迁移学习和半监督学习,引入自适应一致性正则化,综合预训练模型的知识以及半监督中带标签 / 不带标签数据的缺陷信息,提高模型检测性能;将原始样本与增强样本的分类结果作为衡量其对网络贡献性能的指标,并将贡献性能作为权重应用在分类损失上,使模型充分学习小样本下叶轮表面缺陷知识,提高实际工业环境下模型对叶轮表面缺陷检测的泛化能力。试验结果表明,与其他半监督域自适应方法比较,所提方法有效提升叶轮表面缺陷检测精度,不仅为实际工业应用提供了有力的技术支持,也丰富了缺陷检测领域的技术手段,对于推动工业智能化和保障设备安全具有重要的理论价值和实践意义。

Abstract

With the development of industrial intelligence, impeller surface defect detection has played a key role in guaranteeing equipment safety and improving production efficiency. However, the problems of incomplete defect sample information, time-consuming sample labeling, and the poor adaptability of detection models in actual industrial environments constrain improvements in detection accuracy. To solve these problems, a method for impeller surface defect detection is proposed, based on a Fourier-based adaptive consistency regularization (FACR) network. Initially, a source model is pretrained with its parameters held constant. Then, a target model is initialized, following the principles of transfer learning, in which the feature extractor utilizes the parameters of the source model and the parameters of the classifier borrow from the concept of imprinting, thereby leveraging the knowledge of the feature extractor of the source model for informed initialization. An adaptive knowledge consistency strategy is designed to ensure that the features extracted by the target model closely resemble those extracted by the source model. Both labeled and unlabeled data serve as bridges for knowledge transfer. The Kullback-Leibler divergence is employed to measure the dissimilarity in feature extraction between the pretrained and target feature extractors, thereby preventing discrepancies that may lead to negative transfer. In addition, an adaptive distribution consistency strategy is devised by using the structural information contained in the unsupervised data to guide the training of the supervised data, thereby minimizing the gaps between the representations of features from the unsupervised and supervised data. Furthermore, a strategy based on Fourier transformation is explored to address the issue of the inaccurate extraction of defect feature information by existing deep transfer networks. Leveraging the inherent characteristics of the Fourier transformation, where the phase component of the Fourier spectrum retains high-order semantics and the amplitude component contains low-order statistical information, defect images from different domains undergo a Fourier transformation to obtain phase and amplitude information. Linear interpolation is then applied to enhance the amplitude of the sample information, thereby ensuring a phase-invariant exchange of the corresponding regions. This approach directs the focus of the model toward the phase information (i.e., structural details) to enhance feature information and reduce the impacts of background environments on the detection model. Subsequently, the concept of curriculum learning is introduced to dynamically adjust the thresholds. Without introducing new hyperparameters or incurring additional computational costs, fixed thresholds are transformed into real-time adjusted thresholds based on the category difficulty. This efficient utilization of unlabeled data aims to enhance the model performance, improve the classification accuracy for various defect types, and address the challenge of low accuracy in complex defect detection caused by fixed thresholds that do not consider different learning states or difficulties for different categories. Finally, the classification results of the original and enhanced samples are used as metrics with which to measure their respective contributions to the network. These contribution performances are applied as weights to the classification loss, thereby allowing the model to thoroughly learn impeller surface defects from limited samples. This approach addresses the issue of potential discrepancies in the predictions between the original and enhanced images after amplitude interpolation through Fourier transformation. The goal is to enhance the generalizability of the model for impeller surface defect detection in a real industrial environment. ResNet18 is selected as the backbone network to validate the superiority of the proposed model. Under the same experimental conditions and with the number of supervised training samples in the target domain set to 5, 10, and 20, comparisons are made with other mainstream domain adaptation methods. The experimental results demonstrate that, compared with other semi-supervised domain adaptation methods, the proposed method effectively enhances the accuracy of impeller surface defect detection. The method presented in this study can improve the extraction capability of common features in the two domains and thus reduces the requirements for actual impeller surface defect samples and labeling. Simultaneously, using a feature enhancement approach increased the classification accuracy of the model. A favorable detection outcome can be achieved when the defect features exhibit a certain degree of similarity, even for components with different lighting conditions and surface textures. Therefore, this method not only provides strong technical support for practical industrial applications but also enriches the technical means in the field of defect detection, which has important theoretical value and practical significance for guaranteeing equipment safety and promoting industrial intelligence.

关键词

叶轮 /表面缺陷 /半监督 /域自适应 /傅里叶约束 /一致性正则化{{custom_keyword}} /

Key words

impeller /surface defects /semi-supervised /domain adaptive /fourier-based /consistency regularization{{custom_keyword}} /

参考文献

[1] 黄健,郑春厚,章军,等. 基于小样本度量迁移学习的表面缺陷检测[J]. 模式识别与人工智能,2021,34(5):407-414.
HUANG Jian, ZHENG Chunhou, ZHANG Jun, et al.Few-shot metric transfer learning network for surface defect detection[J]. Pattern Recognition and Artificial Intelligence, 2021, 34(5): 407-414. (in Chinese)
[2] FU G, SUN P, ZHU W, et al.A deep-learning-based approach for fast and robust steel surface defects classification[J]. Optics and Lasers in Engineering, 2019, 121: 397-405.
[3] 孙凯. 基于机器学习的带钢表面缺陷分类[D]. 武汉:华中科技大学,2016.
SUN Kai.Machine learning based surface defect classification for steel strips[D]. Wuhan: Huazhong University of Science & Technology, 2016. (in Chinese)
[4] 刘佳,唐鋆磊,林冰,等. 基于HSV(色相-饱和度-明度)与形状特征的涂层锈点图像识别[J]. 中国表面工程,2023,36(4):217-228.
LIU Jia, TANG Junlei, LIN Bing, et al.Rust spot image recognition of coatings based on HSV and shape feature[J]. China Surface Engineering, 2023, 36(4): 217-228. (in Chinese)
[5] 刘泽,王嵬,王平. 钢轨表面缺陷检测机器视觉系统的设计[J]. 电子测量与仪器学报,2010,24(11):1012-1017.
LIU Ze, WANG Wei, WANG Ping.Design of machine vision system for inspection of rail surface defects[J]. Journal of Electronic Measurement and Instrumentation, 2010, 24(11): 1012-1017. (in Chinese)
[6] 何存富,刘溯,刘增华,等. 小波降噪在钢绞线缺陷检测中的应用[J]. 机械工程学报,2008,44(7):118-122.
HE Cunfu, LIU Su, LIU Zenghua, et al.Application of wavelet denoise in defect inspection of steel strands[J]. Chinese Journal of Mechanical Engineering, 2008, 44(7): 118-122. (in Chinese)
[7] LI Y, DU X, WAN F, et al.Rotating machinery fault diagnosis based on convolutional neural network and infrared thermal imaging[J]. Chinese Journal of Aeronautics, 2020, 167(2): 42-53.
[8] ZHANG S, ZHANG Q, GU J, et al.Visual inspection of steel surface defects based on domain adaptation and adaptive convolutional neural network[J]. Mechanical Systems and Signal Processing, 2021, 153: 1-17.
[9] ZHOU S, CHEN Y, ZHANG D, et al.Classification of surface defects on steel sheet using convolutional neural networks[J]. Materiali in Tehnologije, 2017, 51(1): 123-131.
[10] 李可,祁阳,宿磊,等. 基于改进ACGAN的钢表面缺陷视觉检测方法[J]. 机械工程学报,2022,58(24):32-40.
LI Ke, QI Yang, SU Lei, et al.Visual inspection of steel surface defects based on improved auxiliary classification generation adversarial network[J]. Journal of Mechanical Engineering, 2022, 58(24): 32-40. (in Chinese)
[11] ZHANG S, SU L, GU J, et al.Rotating machinery fault detection and diagnosis based on deep domain adaptation: A survey[J]. Chinese Journal of Aeronautics, 2023, 36(1): 45-74.
[12] ROY S, SIAROHIN A, SANGINETO E, et al.Unsupervised domain adaptation using feature-whitening and consensus loss[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 15-20, 2019, Long Beach, CA, USA. IEEE, 2019: 9471-9480.
[13] LOPEZPAZ D, HERNANDEZLOBATO J M, SCHOLKOPF B, et al.Semi-supervised domain adaptation with non-parametric copulas[J]. Neural Information Processing Systems, 2012, 1: 665-673.
[14] JAKUBOVITZ D, RODRIGUES M, GIRYES R.Lautum regularization for semi-supervised transfer learning[C]//2019 IEEE/CVF International Conference on Computer Vision Workshops, October 27-November 2, 2019, Seoul, Korea. IEEE, 2019: 763-767.
[15] LEE D H.Pseudo-label: the simple and efficient semi-supervised learning method for deep neural networks[C]//The 30th International Conference on Machine Learning, June 16-21, 2013, Atlanta, GA, USA. ACM, 2013: 896-901.
[16] ABUDUWEILI A, LI X, SHI H, et al.Adaptive consistency regularization for semi-supervised transfer learning[C]//2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 20-25, 2021, Nashville, TN, USA. IEEE, 2021: 6919-6928.
[17] ZHANG H, CISSE M, DAUPHIN Y N, et al.Mixup: beyond empirical risk minimization[C]//The 6th International Conference on Learning Representations, April 30-May 3, 2018, Vancouver, Canada. OpenReview, 2018: 1-13.
[18] BENGIO Y.Curriculum learning[C]//The 26th International Conference on Machine Learning, June 14-16, 2009, Montreal, Canada. ACM, 2009: 41-48.
[19] 罗东亮,蔡雨萱,杨子豪,等. 工业缺陷检测深度学习方法综述[J]. 中国科学:信息科学,2022,52(6):1002-1039.
LUO Dongliang, CAI Yuxuan, YANG Zihao, et al.Survey on industrial defect detection with deep learning[J]. Scientia Sinica Informationis, 2022, 52(6): 1002-1039. (in Chinese)
[20] GANIN Y, USTINOVA E, AJAKAN H, et al.Domain-adversarial training of neural networks[J]. Journal of Machine Learning Research, 2016, 17(1): 2096-2030.
[21] MOTIIAN S, JONES Q, IRANMANESH S M, et al.Few-shot adversarial domain adaptation[C]//The 31st Conference on Neural Information Processing Systems, December 4-9, 2017, Long Beach, CA, USA. Curran Associates, 2017: 6673-6683.
[22] SAITO K, WATANABE K, USHIKU Y, et al.Maximum classifier discrepancy for unsupervised domain adaptation[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 18-22, 2018, Salt Lake City, UT, USA. IEEE, 2018: 3723-3732.