《计算机应用》唯一官方网站 ›› 2026, Vol. 46 ›› Issue (2): 659-665.DOI: 10.11772/j.issn.1001-9081.2025020243
• 前沿与综合应用 • 上一篇
收稿日期:2025-03-12
修回日期:2025-04-18
接受日期:2025-04-28
发布日期:2025-05-16
出版日期:2026-02-10
通讯作者:
李南莎
作者简介:李海丰(1984—),男,内蒙古通辽人,教授,博士,CCF会员,主要研究方向:机器人环境感知、计算机视觉基金资助:
Haifeng LI1, Wenqiang LIU1, Nansha LI1(
), Zhongcheng GUI2
Received:2025-03-12
Revised:2025-04-18
Accepted:2025-04-28
Online:2025-05-16
Published:2026-02-10
Contact:
Nansha LI
About author:LI Haifeng, born in 1984, Ph. D., professor. His research interests include robot environmental perception, computer vision.Supported by:摘要:
针对机场跑道探地雷达(GPR)数据中的复杂背景杂波和层间强反射干扰信号的问题,提出一种基于改进U-Net的深度学习杂波抑制算法。该算法在U-Net的跳跃连接处引入细节增强模块DE-Conv,从而增强网络对多尺度浅层特征中目标信号细节的捕捉能力;同时,采用含杂波-无杂波图像对计算特征-像素双级融合损失函数优化训练过程。具体地,通过共享权重编码器提取的含杂波与无杂波数据的高维特征,计算特征级别损失来指导编码器的训练,并使用解码器输出图像与对应的无杂波仿真图像计算像素级别损失以优化解码器性能。实验结果表明,在合成数据集上,所提算法的峰值信噪比(PSNR)和结构相似度(SSIM)分别达到37.114 7 dB和0.999 8;而在真实机场跑道数据集上,所提算法的平均信杂比(SCR)和改善系数(IF)分别为8.28 dB和5.90 dB,以上4种指标相较于基准模型的数据分别提升了0.952 8 dB、0.000 4、6.58 dB和5.32 dB。与鲁棒非负矩阵分解(RNMF)、鲁棒主成分分析(RPCA)及同样基于深度学习的基于U-Net改进的杂波去除神经网络(CR-Net)相比,所提算法在杂波抑制效果和计算效率上均表现出优势。同时,大量的消融实验结果验证了细节增强模块和特征-像素双级损失函数对杂波去除和目标信号恢复的有效性。
中图分类号:
李海丰, 刘文强, 李南莎, 桂仲成. 面向机场跑道的探地雷达杂波抑制算法[J]. 计算机应用, 2026, 46(2): 659-665.
Haifeng LI, Wenqiang LIU, Nansha LI, Zhongcheng GUI. Ground penetrating radar clutter suppression algorithm for airport runways[J]. Journal of Computer Applications, 2026, 46(2): 659-665.
| 背景 | 仿真 | 合成数据 | 真实数据 | |
|---|---|---|---|---|
| 训练 | 测试 | 测试 | ||
| 83 | 1 628 | 1 466 | 162 | 232 |
表1 数据集样本的分布
Tab. 1 Dataset sample distribution
| 背景 | 仿真 | 合成数据 | 真实数据 | |
|---|---|---|---|---|
| 训练 | 测试 | 测试 | ||
| 83 | 1 628 | 1 466 | 162 | 232 |
| 算法 | PSNR/dB(↑) | SSIM(↑) | MAE(↓) | MSE(↓) |
|---|---|---|---|---|
| MS[ | 30.445 8 | 0.978 7 | 107.877 7 | 59.148 5 |
| SVD[ | 31.696 3 | 0.968 6 | 106.506 0 | 44.824 5 |
| RNMF[ | 30.085 9 | 0.989 3 | 120.939 4 | 64.820 3 |
| RPCA[ | 32.857 1 | 0.547 9 | 101.022 1 | 34.653 0 |
| CR-Net[ | 36.772 9 | 0.999 4 | 98.319 3 | 14.140 3 |
| 本文算法 | 37.114 7 | 0.999 8 | 96.101 9 | 13.111 7 |
表2 不同算法在合成数据上的背景抑制性能比较
Tab. 2 Comparison of background suppression performance of different algorithms on synthetic data
| 算法 | PSNR/dB(↑) | SSIM(↑) | MAE(↓) | MSE(↓) |
|---|---|---|---|---|
| MS[ | 30.445 8 | 0.978 7 | 107.877 7 | 59.148 5 |
| SVD[ | 31.696 3 | 0.968 6 | 106.506 0 | 44.824 5 |
| RNMF[ | 30.085 9 | 0.989 3 | 120.939 4 | 64.820 3 |
| RPCA[ | 32.857 1 | 0.547 9 | 101.022 1 | 34.653 0 |
| CR-Net[ | 36.772 9 | 0.999 4 | 98.319 3 | 14.140 3 |
| 本文算法 | 37.114 7 | 0.999 8 | 96.101 9 | 13.111 7 |
| 算法 | SCR | IF | 推理时间/s |
|---|---|---|---|
| MS | 1.77 | 0.32 | 0.011 |
| SVD | 2.13 | 1.00 | 0.244 |
| RNMF | 1.76 | 0.18 | 10.457 |
| RPCA | 3.81 | 2.93 | 1.340 |
| CR-Net | 6.02 | 4.18 | 0.238 |
| 本文算法 | 8.28 | 5.90 | 0.231 |
表3 不同算法在真实数据上的背景抑制性能比较
Tab. 3 Comparison of background suppression performance of different algorithms on real data
| 算法 | SCR | IF | 推理时间/s |
|---|---|---|---|
| MS | 1.77 | 0.32 | 0.011 |
| SVD | 2.13 | 1.00 | 0.244 |
| RNMF | 1.76 | 0.18 | 10.457 |
| RPCA | 3.81 | 2.93 | 1.340 |
| CR-Net | 6.02 | 4.18 | 0.238 |
| 本文算法 | 8.28 | 5.90 | 0.231 |
| 实验序号 | DE-Conv | 双流损失 | 合成数据 | 真实数据 | |||||
|---|---|---|---|---|---|---|---|---|---|
| P1+P2+P3 | P4 | PSNR/dB(↑) | SSIM(↑) | MAE(↓) | MSE(↓) | 平均SCR/dB(↑) | 平均IF/dB(↑) | ||
| 1 | × | × | × | 36.161 9 | 0.999 4 | 109.367 0 | 16.230 3 | 1.70 | 0.58 |
| 2 | √ | √ | × | 36.644 7 | 0.999 5 | 97.018 8 | 14.614 0 | 2.33 | 0.70 |
| 3 | × | × | √ | 36.774 9 | 0.999 7 | 100.605 6 | 14.175 7 | 3.05 | 1.11 |
| 4 | √ | × | √ | 36.892 1 | 0.999 8 | 98.356 2 | 13.139 0 | 8.02 | 5.67 |
| 5 | √ | √ | √ | 37.114 7 | 0.999 8 | 96.101 9 | 13.111 7 | 8.28 | 5.90 |
表4 DE-Conv模块和双级损失模块的消融实验结果
Tab. 4 Ablation experiment results of DE-Conv module and dual-level loss module
| 实验序号 | DE-Conv | 双流损失 | 合成数据 | 真实数据 | |||||
|---|---|---|---|---|---|---|---|---|---|
| P1+P2+P3 | P4 | PSNR/dB(↑) | SSIM(↑) | MAE(↓) | MSE(↓) | 平均SCR/dB(↑) | 平均IF/dB(↑) | ||
| 1 | × | × | × | 36.161 9 | 0.999 4 | 109.367 0 | 16.230 3 | 1.70 | 0.58 |
| 2 | √ | √ | × | 36.644 7 | 0.999 5 | 97.018 8 | 14.614 0 | 2.33 | 0.70 |
| 3 | × | × | √ | 36.774 9 | 0.999 7 | 100.605 6 | 14.175 7 | 3.05 | 1.11 |
| 4 | √ | × | √ | 36.892 1 | 0.999 8 | 98.356 2 | 13.139 0 | 8.02 | 5.67 |
| 5 | √ | √ | √ | 37.114 7 | 0.999 8 | 96.101 9 | 13.111 7 | 8.28 | 5.90 |
| [1] | KOOHMISHI M, KAEWUNRUEN S, CHANG L, et al. Advancing railway track health monitoring: integrating GPR, InSAR and machine learning for enhanced asset management[J]. Automation in Construction, 2024, 162: No.105378. |
| [2] | 杜翠,张千里,刘杰. 基于HCA与KAZE的铁路路基GPR图像配准算法[J]. 计算机工程, 2018, 44(3): 264-269, 274. |
| DU C, ZHANG Q L, LIU J. Railway ballast GPR image registration algorithm based on HCA and KAZE[J]. Computer Engineering, 2018, 44(3): 264-269, 274. | |
| [3] | 黄晓惠. 透水混凝土整体路面施工质量控制与检测技术研究[D]. 绵阳:西南科技大学, 2021: 47-52. |
| HUANG X H. Research on construction quality control and detection technology for pervious concrete monolithic pavement[D]. Mianyang: Southwest University of Science and Technology, 2021: 47-52. | |
| [4] | 郭云飞. 机场水泥混凝土道面脱空注浆修复评价研究[D]. 郑州:郑州大学, 2021: 46-56. |
| GUO Y F. Research on evaluation of void grouting repair for airport cement concrete pavement[D]. Zhengzhou: Zhengzhou University, 2021: 46-56. | |
| [5] | LIU K, GAO Z, CHEN F, et al. Application of airport pavement structure safety detection and intelligent recognition technology[C]// Proceedings of the 4th International Symposium on Traffic Transportation and Civil Architecture. Piscataway: IEEE, 2021: 143-146. |
| [6] | ZHANG Y, TONG Z, SHE X, et al. SWC-Net and multi-phase heterogeneous FDTD model for void detection underneath airport pavement slab[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(12): 20698-20714. |
| [7] | HU Q F, ZHANG Z H, WANG F, et al. 3D GPR detection and case analysis of collapse hazards for an international airport runway[J]. Journal of Physics: Conference Series, 2024, 2887: No.012026. |
| [8] | LUO W, LEE Y H, JIAN X, et al. A new method for GPR clutter suppression based on stationary graph signals processing[J]. IEEE Transactions on Geoscience and Remote Sensing, 2025, 63: No.4500112. |
| [9] | HOU F, FANG M, LUO T X, et al. Dual-task GPR method: improved generative adversarial clutter suppression network and adaptive target localization algorithm in GPR image[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: No.5108313. |
| [10] | TANG X S, YANG F, QIAO X, et al. A ground-penetrating radar clutter suppression algorithm integrating signal processing and image fusion[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: No.5936618. |
| [11] | SOLIMENE R, CUCCARO A, DELL’AVERSANO A, et al. Ground clutter removal in GPR surveys[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(3): 792-798. |
| [12] | BENEDETTO A, TOSTI F, BIANCHINI CIAMPOLI L, et al. An overview of ground-penetrating radar signal processing techniques for road inspections[J]. Signal Processing, 2017, 132: 201-209. |
| [13] | KUMLU D, ERER I. Improved clutter removal in GPR by robust nonnegative matrix factorization[J]. IEEE Geoscience and Remote Sensing Letters, 2020, 17(6): 958-962. |
| [14] | ABUJARAD F, NADIM G, OMAR A. Clutter reduction and detection of landmine objects in ground penetrating radar data using Singular Value Decomposition (SVD)[C]// Proceedings of the 3rd International Workshop on Advanced Ground Penetrating Radar. Piscataway: IEEE, 2005: 37-42. |
| [15] | CHEN G, FU L, CHEN K, et al. Adaptive ground clutter reduction in ground-penetrating radar data based on principal component analysis[J]. IEEE Transactions on Geoscience and Remote Sensing, 2019, 57(6): 3271-3282. |
| [16] | SUN H H, CHENG W, FAN Z. Learning to remove clutter in real-world GPR images using hybrid data[J]. IEEE Transactions on Geoscience and Remote Sensing, 2022, 60: No.5113714. |
| [17] | SONG X, XIANG D, ZHOU K, et al. Improving RPCA-based clutter suppression in GPR detection of antipersonnel mines[J]. IEEE Geoscience and Remote Sensing Letters, 2017, 14(8): 1338-1342. |
| [18] | 雷文太,毛凌青,庞泽邦,等. DR-GAN:一种无监督学习的探地雷达杂波抑制方法[J]. 电子与信息学报, 2023, 45(10): 3776-3785. |
| LEI W T, MAO L Q, PANG Z B, et al. DR-GAN: an unsupervised learning approach to clutter suppression for ground penetrating radar[J]. Journal of Electronics and Information Technology, 2023, 45(10): 3776-3785. | |
| [19] | 李海丰,张凡,朴敏楠,等. 基于通道和空间注意力的机场道面地下目标自动检测[J]. 计算机应用, 2023, 43(3): 930-935. |
| LI H F, ZHANG F, PIAO M N, et al. Automatic detection of targets under airport pavement based on channel and spatial attention[J]. Journal of Computer Applications, 2023, 43(3): 930-935. | |
| [20] | LI N, WU R, LI H, et al. M2FNet: multi-modal fusion network for airport runway subsurface defect detection using GPR data[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: No.5108816. |
| [21] | NI Z K, SHI C, PAN J, et al. Declutter-GAN: GPR B-scan data clutter removal using conditional generative adversarial nets[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19: No.4023105. |
| [22] | RONNEBERGER O, FISCHER P, BROX T. U-Net: convolutional networks for biomedical image segmentation[C]// Proceedings of the 2015 Medical Image Computing and Computer-Assisted Intervention, LNCS 9351. Cham: Springer, 2015: 234-241. |
| [23] | PANDA S L, SAHOO U K, MAITI S, et al. An attention U-Net based improved clutter suppression in GPR images[J]. IEEE Transactions on Instrumentation and Measurement, 2024, 73: No. 8502511. |
| [24] | LIU G, LAN S, ZHANG T, et al. SAGAN: skip-attention GAN for anomaly detection[C]// Proceedings of the 2021 IEEE International Conference on Image Processing. Piscataway: IEEE, 2021: 2468-2472. |
| [25] | WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[C]// Proceedings of the 2018 European Conference on Computer Vision, LNCS 11211. Cham: Springer, 2018: 3-19. |
| [26] | YANG C, LAN S, HUANG W, et al. A Transformer-based GAN for anomaly detection[C]// Proceedings of the 2022 International Conference on Artificial Neural Networks, LNCS 13530. Cham: Springer, 2022: 345-357. |
| [27] | PU B, LAN S, WANG W, et al. GanNeXt: a new convolutional GAN for anomaly detection[C]// Proceedings of the 2023 International Conference on Artificial Neural Networks, LNCS 14256. Cham: Springer, 2023: 39-49. |
| [28] | CHEN Z, HE Z, LU Z M. DEA-Net: single image dehazing based on detail-enhanced convolution and content-guided attention[J]. IEEE Transactions on Image Processing, 2024, 33: 1002-1015. |
| [1] | 李海丰, 张凡, 朴敏楠, 王怀超, 李南莎, 桂仲成. 基于通道和空间注意力的机场道面地下目标自动检测[J]. 《计算机应用》唯一官方网站, 2023, 43(3): 930-935. |
| [2] | 李海丰, 赵碧帆, 侯谨毅, 王怀超, 桂仲成. 基于自适应双阈值的地下目标自动检测算法[J]. 《计算机应用》唯一官方网站, 2022, 42(4): 1275-1283. |
| [3] | 肖磊, 熊秀娟, 陈菲, 陈波. 超声血流成像中基于动态域的回归和奇异值分解的杂波抑制方法[J]. 计算机应用, 2015, 35(1): 265-269. |
| [4] | 吴静 王洪 汪学刚. 机场跑道异物监测雷达的杂波图恒虚警率检测[J]. 计算机应用, 2013, 33(11): 3288-3290. |
| [5] | 汪成曦 刘以安 张强. 改进的最小均方自适应滤波算法[J]. 计算机应用, 2012, 32(07): 2078-2081. |
| [6] | 丁建立 李晓丽 李全福. 基于改进蚁群协同算法的枢纽机场场面滑行道优化调度模型[J]. 计算机应用, 2010, 30(4): 1000-1003. |
| [7] | 杨莘,陈淑珍,唐中柱. 探地雷达管道目标图像的识别[J]. 计算机应用, 2005, 25(05): 1209-1211. |
| 阅读次数 | ||||||
|
全文 |
|
|||||
|
摘要 |
|
|||||