

浏览全部资源
扫码关注微信
哈尔滨理工大学 黑龙江省激光光谱技术及应用重点实验室,黑龙江 哈尔滨 150080
Received:26 May 2026,
Revised:2026-07-07,
Online First:14 August 2026,
Published:10 August 2026
移动端阅览
敬天池,孙晓明,郝亮等.基于双重频域增强与动态核引导的DFD-DETR钢材表面缺陷检测[J].光学精密工程,2026,34(15):2422-2437.
JING Tianchi,SUN Xiaoming,HAO Liang,et al.Steel surface defect detection based on DFD-DETR with dual-frequency enhancement and dynamic kernel guidance[J].Optics and Precision Engineering,2026,34(15):2422-2437.
敬天池,孙晓明,郝亮等.基于双重频域增强与动态核引导的DFD-DETR钢材表面缺陷检测[J].光学精密工程,2026,34(15):2422-2437. DOI: 10.37188/OPE.20263415.2422. CSTR: 32169.14.OPE.20263415.2422.
JING Tianchi,SUN Xiaoming,HAO Liang,et al.Steel surface defect detection based on DFD-DETR with dual-frequency enhancement and dynamic kernel guidance[J].Optics and Precision Engineering,2026,34(15):2422-2437. DOI: 10.37188/OPE.20263415.2422. CSTR: 32169.14.OPE.20263415.2422.
针对钢铁表面缺陷检测中复杂背景干扰导致的低对比度缺陷特征难以表征,以及跨尺度特征融合过程中局部细节易丢失的问题,提出一种融合双重频域增强与动态核引导的改进型检测框架 DFD-DETR。首先,针对轧制纹理等低频背景对缺陷边缘特征的干扰,设计了双重频域增强模块(Dual-Frequency Domain Enhancement, DFE)。该模块由前置高通滤波与后置频谱门控两部分组成,前者通过固定高通响应抑制轧制纹理等低频背景信息,后者利用可学习频谱门控对融合后的频域特征进行自适应重校准,从而增强微弱缺陷的边缘响应。随后,为在一定程度上克服传统特征融合过程中浅层细节被深层语义覆盖的困难,提出动态核引导模块(Dynamic Kernel Guidance, DKG)。利用深层语义特征生成空间自适应动态卷积核,对浅层高分辨率特征进行局部引导融合,从而增强多尺度缺陷的结构表征能力。在NEU-DET 与 GC10-DET 数据集上的实验结果表明,本文方法的mAP@0.5分别达到80.8%与73.4%,相比基线RF-DETR,所提出方法在复杂背景下的低对比度及多尺度缺陷检测任务中取得了更优的检测结果,具有较好的检测性能与一定的泛化能力。
Steel surface defect detection is challenged by complex industrial backgrounds, which weaken low-contrast defect representations, as well as by the loss of local structural details during cross-scale feature fusion. To address these issues, an improved detection framework, termed DFD-DETR, is proposed, integrating a dual-frequency-domain enhancement strategy with a dynamic kernel guidance mechanism. Specifically, a Dual-Frequency Domain Enhancement (DFE) module is designed to suppress low-frequency background interference caused by rolling textures while enhancing the high-frequency edge responses of subtle defects. The DFE module consists of a pre-stage high-pass filtering module and a post-stage spectral gating module. In the pre-stage, low-frequency background components, such as rolling textures, are suppressed by a fixed high-pass frequency response, whereas in the post-stage, the fused frequency-domain features are adaptively recalibrated through a learnable spectral gating mechanism, thereby strengthening the edge responses of subtle defects. To further mitigate the loss of fine-grained information during multi-scale feature interaction, a Dynamic Kernel Guidance (DKG) module is introduced. Spatially adaptive dynamic convolution kernels are generated from high-level semantic features, and locally guided aggregation is performed on high-resolution shallow features, thereby enhancing structural representation across multiple defect scales. Extensive experiments on the NEU-DET and GC10-DET datasets demonstrate that the proposed method achieves mAP@0.5 scores of 80.8% and 73.4%, respectively. Compared with the baseline RF-DETR, the proposed framework exhibits superior performance in low-contrast and multi-scale defect detection under complex backgrounds, indicating improved detection capability and generalization performance.
唐东林 , 杨洲 , 程衡 , 等 . 浅层卷积神经网络融合Transformer的金属缺陷图像识别方法 [J]. 中国机械工程 , 2022 , 33 ( 19 ): 2298 - 2305, 2316 .
TANG D L , YANG ZH , CHENG H , et al . Metal defect image recognition method based on shallow CNN fusion transformer [J]. China Mechanical Engineering , 2022 , 33 ( 19 ): 2298 - 2305, 2316 . (in Chinese)
邓志鹏 , 何施茗 , 杨根 , 等 . 基于深度学习的纹理表面缺陷检测方法综述 [J]. 计算机集成制造系统 , 2025 , 31 ( 3 ): 721 - 745 .
DENG ZH P , HE SH M , YANG G , et al . Survey of texture surface defect detection method based on deep learning [J]. Computer Integrated Manufacturing Systems , 2025 , 31 ( 3 ): 721 - 745 . (in Chinese)
李少波 , 杨静 , 王铮 , 等 . 缺陷检测技术的发展与应用研究综述 [J]. 自动化学报 , 2020 , 46 ( 11 ): 2319 - 2336 . doi: 10.16383/j.aas.c180538 http://dx.doi.org/10.16383/j.aas.c180538
LI SH B , YANG J , WANG ZH , et al . Review of development and application of defect detection technology [J]. Acta Automatica Sinica , 2020 , 46 ( 11 ): 2319 - 2336 . (in Chinese) . doi: 10.16383/j.aas.c180538 http://dx.doi.org/10.16383/j.aas.c180538
REDMON J , DIVVALA S , GIRSHICK R , et al . You only look once: unified, real-time object detection [C]. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 27-30,2016 , Las Vegas, NV, USA. IEEE , 2016 : 779 - 788 . doi: 10.1109/cvpr.2016.91 http://dx.doi.org/10.1109/cvpr.2016.91
CARION N , MASSA F , SYNNAEVE G , et al . End-to-end object detection with transformers [EB/OL]. 2020 : arXiv : 2005 . 12872 . https://arxiv.org/abs/2005.12872 https://arxiv.org/abs/2005.12872 . doi: 10.1007/978-3-030-58452-8_13 http://dx.doi.org/10.1007/978-3-030-58452-8_13
ZHU X Z , SU W J , LU L W , et al . Deformable DETR: deformable transformers for end-to-end object detection [EB/OL]. 2020 : arXiv : 2010 . 04159 . https://arxiv.org/abs/2010.04159 https://arxiv.org/abs/2010.04159 .
LIN T Y , DOLLÁR P , GIRSHICK R , et al . Feature pyramid networks for object detection [C]. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 21-26,2017 , Honolulu, HI, USA. IEEE , 2017 : 936 - 944 . doi: 10.1109/cvpr.2017.106 http://dx.doi.org/10.1109/cvpr.2017.106
LIU S , QI L , QIN H F , et al . Path aggregation network for instance segmentation [C]. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition . 18-23,2018 , Salt Lake City, UT, USA . IEEE , 2018 : 8759 - 8768 . doi: 10.1109/cvpr.2018.00913 http://dx.doi.org/10.1109/cvpr.2018.00913
XIAO H Y , ZHANG W Y , ZUO L , et al . WFF-Net: Trainable weight feature fusion convolutional neural networks for surface defect detection [J]. Advanced Engineering Informatics , 2025 , 64 : 103073 . doi: 10.1016/j.aei.2024.103073 http://dx.doi.org/10.1016/j.aei.2024.103073
郭伟 , 孟巧巧 , 金海波 , 等 . 多尺度交互与动态协同的钢表面缺陷检测 [J]. 计算机工程 , 2025 : 1 - 14 .
GUO W , MENG Q Q , JIN H B , et al . Multi-scale Interaction and dynamic collaboration for steel surface defect detection [J]. Computer Engineering , 2025 : 1 - 14 . (in Chinese)
张晓檬 , 黄卫阳 , 吕文丹 , 等 . 基于GAB-YOLO的钢材表面缺陷检测方法 [J]. 计算机测量与控制 , 2026 : 1 - 10 .
ZHANG X M , HUANG W Y , LYU W D , et al . Steel surface defect detection method based on GAB-YOLO [J]. Computer Measurement & Control , 2026 : 1 - 10 . (in Chinese)
彭有有 , 李维刚 , 胡佳琪 . 基于DMI-YOLO算法的带钢表面缺陷检测 [J]. 钢铁研究学报 , 2026 , 38 ( 3 ): 373 - 383 .
PENG Y Y , LI W G , HU J Q . Strip steel surface defect detection based on DMI-YOLO algorithm [J]. Journal of Iron and Steel Research , 2026 , 38 ( 3 ): 373 - 383 . (in Chinese)
LIU G H , CHU M X , GONG R F , et al . Global attention module and cascade fusion network for steel surface defect detection [J]. Pattern Recognition , 2025 , 158 : 110979 . doi: 10.1016/j.patcog.2024.110979 http://dx.doi.org/10.1016/j.patcog.2024.110979
LI J , LUO W , et al . PCBA pin solder defect detection method based on improved RT-DETR [J]. Chinese Journal of Liquid Crystals and Displays , 2025 , 40 ( 10 ): 1532 - 1544 . doi: 10.37188/cjlcd.2025-0143 http://dx.doi.org/10.37188/cjlcd.2025-0143
ROBINSON I , ROBICHEAUX P , POPOV M , et al . RF-DETR: neural architecture search for real-time detection transformers [EB/OL]. 2025 : arXiv : 2511 . 09554 . https://arxiv.org/abs/2511.09554 https://arxiv.org/abs/2511.09554 .
PEI X , HUANG Y D , SU W J , et al . FFTFormer: a spatial-frequency noise aware CNN-Transformer for low light image enhancement [J]. Knowledge-Based Systems , 2025 , 314 : 113055 . doi: 10.1016/j.knosys.2025.113055 http://dx.doi.org/10.1016/j.knosys.2025.113055
LIU S L , ZENG Z Y , REN T H , et al . Grounding DINO: marrying DINO with grounded pre-training for open-set object detection [C]. Computer Vision-ECCV 2024. Cham : Springer , 2025 : 38 - 55 . doi: 10.1007/978-3-031-72970-6_3 http://dx.doi.org/10.1007/978-3-031-72970-6_3
CHEN Y P , DAI X Y , LIU M C , et al . Dynamic convolution: attention over convolution kernels [C]. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 13 - 19 , 2020 . Seattle, WA, USA. IEEE , 2020 : 11027 - 11036 .
WANG J Q , CHEN K , XU R , et al . CARAFE: content-aware ReAssembly of FEatures [C]. 2019 IEEE/CVF International Conference on Computer Vision (ICCV). October 27-November 2 , 2019 . Seoul, Korea. IEEE , 2019: 3007 - 3016 . doi: 10.1109/iccv.2019.00310 http://dx.doi.org/10.1109/iccv.2019.00310
HE Y , SONG K C , MENG Q G , et al . An end-to-end steel surface defect detection approach via fusing multiple hierarchical features [J]. IEEE Transactions on Instrumentation and Measurement , 2020 , 69 ( 4 ): 1493 - 1504 . doi: 10.1109/tim.2019.2915404 http://dx.doi.org/10.1109/tim.2019.2915404
LV X M , DUAN F J , JIANG J J , et al . Deep metallic surface defect detection: the new benchmark and detection network [J]. Sensors , 2020 , 20 ( 6 ): 1562 . doi: 10.3390/s20061562 http://dx.doi.org/10.3390/s20061562
REN Q R , GENG J H , LI J Y . Slighter Faster R-CNN for real-time detection of steel strip surface defects [C]. 2018 Chinese Automation Congress (CAC). 30-2,2018 , Xi’an, China. IEEE , 2019 : 2173 - 2178 . doi: 10.1109/cac.2018.8623407 http://dx.doi.org/10.1109/cac.2018.8623407
LIU W , ANGUELOV D , ERHAN D , et al . SSD : Single Shot MultiBox Detector [M]. Computer Vision-ECCV 2016. Cham : Springer International Publishing , 2016 : 21 - 37 . doi: 10.1007/978-3-319-46448-0_2 http://dx.doi.org/10.1007/978-3-319-46448-0_2
VARGHESE R , M S . YOLOv8: a novel object detection algorithm with enhanced performance and robustness [C]. 2024 International Conference on Advances in Data Engineering and Intelligent Computing Systems (ADICS). 18-19,2024 , Chennai, India. IEEE , 2024 : 1 - 6 . doi: 10.1109/adics58448.2024.10533619 http://dx.doi.org/10.1109/adics58448.2024.10533619
KHANAM R , HUSSAIN M . YOLOv11: an overview of the key architectural enhancements [EB/OL]. 2024 : arXiv : 2410 . 17725 . https://arxiv.org/abs/2410.17725 https://arxiv.org/abs/2410.17725 . doi: 10.3390/solar5010006 http://dx.doi.org/10.3390/solar5010006
WANG C L , XU K . SDF-YOLO: an efficient neural network for detecting surface defects in steel strips [J]. Journal of Physics: Conference Series , 2025 , 2990 ( 1 ): 012024 . doi: 10.1088/1742-6596/2990/1/012024 http://dx.doi.org/10.1088/1742-6596/2990/1/012024
梁礼明 , 龙鹏威 , 冯耀 , 等 . 改进轻量化VTG-YOLOv7-tiny的钢材表面缺陷检测 [J]. 光学 精密工程 , 2024 , 32 ( 8 ): 1227 - 1240 . doi: 10.37188/ope.20243208.1227 http://dx.doi.org/10.37188/ope.20243208.1227
LIANG L M , LONG P W , FENG Y , et al . Improving the lightweight VTG-YOLOv7-tiny for steel surface defect detection [J]. Optics and Precision Engineering , 2024 , 32 ( 8 ): 1227 - 1240 . (in Chinese) . doi: 10.37188/ope.20243208.1227 http://dx.doi.org/10.37188/ope.20243208.1227
张新君 , 赵志闯 . 基于UAL-YOLO的钢材表面缺陷检测算法 [J]. 电子测量与仪器学报 , 2026 : 1 - 12 .
ZHANG X J , ZHAO ZH CH . Steel surface defect detection algorithm based on UAL-YOLO [J]. Journal of Electronic Measurement and Instrumentation , 2026 : 1 - 12 . (in Chinese)
XIA Y K , LU Y , JIANG X H , et al . Enhanced multiscale attentional feature fusion model for defect detection on steel surfaces [J]. Pattern Recognition Letters , 2025 , 188 ( C ): 15 - 21 . doi: 10.1016/j.patrec.2024.11.024 http://dx.doi.org/10.1016/j.patrec.2024.11.024
0
Views
20
下载量
0
CSCD
Publicity Resources
Related Articles
Related Author
Related Institution
京公网安备11010802024621