桥梁伸缩缝作为桥梁结构的关键组件，其健康状态直接关系到桥梁的安全性。桥梁伸缩缝损伤可分为表面损伤以及结构损伤两类：表面损伤主要表现为伸缩缝锚固带裂缝，结构损伤主要表现为伸缩缝位移控制弹簧老化以及约束带松动。与传感器监测方法相比，计算机视觉监测方法具有维护成本低的优势，展现了良好的应用潜力。然而，由于桥梁伸缩缝表面损伤具有裂缝尺寸微小、背景噪声干扰显著的特点，使得目标检测算法检测精度较低。同时，桥梁伸缩缝结构损伤中的异型钢存在边缘模糊、纹理信息弱的特点，使得双目视觉算法测距误差较大。因此，针对以上问题，本文主要研究了基于计算机视觉的桥梁伸缩缝健康监测方法，具体研究内容如下：

针对桥梁伸缩缝表面损伤存在的裂缝尺寸微小、背景噪声干扰显著的特点，提出了一种基于YOLOv5-DAMF的目标检测模型。该模型构建Dense Unit密集连接模块替代原Res Unit模块，每次卷积运算都将先前所有层输出的特征作为该阶段所需的输入，使网络能够更有效地学习裂缝特征。在CBS模块后添加CBAM注意力机制，通过在通道和空间两个维度上对特征图进行加权，减少背景噪声的干扰。提出多尺度特征融合策略并增设小目标检测层，通过整合不同尺度的特征信息，结合低层级细节与高层级语义，增强了模型对微小裂缝特征的捕捉能力。实验表明，相较于原模型，YOLOv5-DAMF的mAP@0.5指标提升了4%达到了86.2%，可以有效地监测桥梁伸缩缝表面损伤。

针对桥梁伸缩缝结构损伤中异型钢存在的边缘模糊以及纹理信息弱的特点，提出了一种基于LAP-SGBM-WLS的双目测距算法。该算法构建拉普拉斯二阶微分卷积核，计算图像中每个像素点的二阶导数来突出像素变化较大的区域，有效提高视差图中异型钢边缘的区分度。结合加权最小二乘滤波，通过最小化加权残差平方和来估算图像中的真实像素值，提升弱纹理区域像素匹配的精准度。实验表明，相较于原算法，LAP-SGBM-WLS算法将测距相对误差从11.32%降低至7.54%，为伸缩缝位移控制弹簧老化、约束带松动等结构损伤的量化评估提供了数据支持。

为了全方位监测桥梁伸缩缝的健康状态，本文开发了一套桥梁伸缩缝健康监测系统。该系统针对表面损伤，集成了基于YOLOv5-DAMF的目标检测算法，用于锚固带裂缝的检测；针对结构损伤，集成了基于LAP-SGBM-WLS的立体匹配算法，用于异型钢伸缩值的测量。系统采用多线程架构和SQLite数据库，支持批量图像处理与实时监测，并通过可视化界面实现参数动态调整与结果交互。

Bridge expansion joints, critical to bridge safety and durability, are prone to surface and structural damage. Surface damage involves anchor belt cracks, while structural damage includes displacement control spring aging and restraint belt loosening. Computer vision-based monitoring, with lower maintenance costs than sensor-based methods, shows potential but faces challenges: tiny cracks with significant background noise lower detection algorithm accuracy, and weak texture details in deformed steel make stereo vision distance measurement error-prone. This study explores computer vision-based health monitoring methods for bridge expansion joints, with three main focuses:

A YOLOv5-DAMF object detection model is proposed for surface damage. Dense Unit replaces the original Res Unit module, allowing the network to learn crack features more effectively. CBAM attention mechanism is added after the CBS module to reduce background noise. A multi-scale feature fusion strategy and a small target detection layer are proposed to enhance the model's ability to capture tiny crack features. Experiments show that compared to YOLOv5, the mAP@0.5 of YOLOv5-DAMF improves by 4% to 86.2%, effectively detecting surface damage.

A LAP-SGBM-WLS stereo matching algorithm is introduced for structural damage. A Laplacian second-order differential convolution kernel highlights areas of significant pixel change, improving the distinction of deformed steel edges in disparity maps.Weighted least squares filtering estimates true pixel values, enhancing matching accuracy in weak texture regions. Experiments indicate that compared to SGBM, LAP-SGBM-WLS reduces relative distance error from 11.32% to 7.54%,providing precise data for structural damage assessment.

A comprehensive bridge expansion joint health monitoring system is developed. It integrates the YOLOv5-DAMF algorithm for crack detection and the LAP-SGBM-WLS algorithm for deformable steel displacement measurement. With a multi-threaded architecture and SQLite database, the system supports batch image processing and real-time monitoring, and a visual interface enables dynamic parameter adjustment and result interaction.

[1]谈珂威.公路桥梁的伸缩缝施工技术[J].价值工程,2024,43(31):91-93.

[2]丁泉,陈朋,章伟,等.大跨径桥梁模数式伸缩缝病害及其控制策略探讨[J].公路,2024,69(05):451-456.

[3]张大卫.公路桥梁伸缩缝锚固区的破坏机理及材料应用探讨[J].交通科技与管理,2024,5(20):143-145.

[4]杨志杰.桥梁梳齿板式伸缩缝设计及施工控制要点[J].城市道桥与防洪,2025,(02):295-298+302.

[5]郑重.公路桥梁模数式伸缩缝常见病害及防治措施[C]//中国公路学会养护与管理分会,重庆高速公路集团有限公司,招商局重庆交通科研设计院有限公司.中国公路学会养护与管理分会第十三届学术年会论文集.中交瑞通路桥养护科技有限公司,2024:444-450.

[6]高珂华.桥梁伸缩缝损坏形式及其影响因素分析[J].交通世界,2018,(Z2):190-191.

[7]袁博,张春雨.公路桥梁异型钢单缝式及模数式伸缩装置病害原因分析及防治措施[J].公路交通科技(应用技术版),2020,16(06):248-251.

[8]王稷良,马淑梅,吕志刚.桥梁伸缩缝锚固区破坏机理及材料应用现状[J].建材世界,2024,45(01):87-91.

[9]李挺.桥梁伸缩缝常见病害及处理、预防措施[J].运输经理世界,2024,(28):142-144.

[10]叶寿瑛.城市桥梁伸缩缝典型病害分析与处理措施[J].福建建材,2021,(11):65-67.

[11]陈宇鹏.普通公路桥梁伸缩缝的病害及养护措施[J].城市建设理论研究(电子版),2024,(02):154-156.

[12]陶斌.大位移模数式桥梁伸缩装置病害分析与预防[J].交通世界,2022,(08):148-150.

[13]顾永标.高速公路桥梁伸缩缝的主要病害及养护措施探讨[J].建筑机械,2023,(08):46-49.

[14]王辉.高速公路桥梁伸缩缝常见问题及防治措施[J].上海建设科技,2023,(03):77-79.

[15]袁博,张春雨.公路桥梁异型钢单缝式及模数式伸缩装置病害原因分析及防治措施[J].公路交通科技(应用技术版),2020,16(06):248-251.

[16]董浩,关长禄.公路桥梁异型钢单缝式伸缩装置常见问题分析[J].中外公路,2017,37(01):152-157.

[17]马宏伟.大位移公路桥梁伸缩缝损伤养护施工技术[J].科学技术创新,2021,(36):112-114.

[18]李微.基于贝叶斯线性模型的桥梁伸缩缝损伤评估及可靠性研究[D].石家庄铁道大学,2021.

[19]刘家齐.高速公路桥梁伸缩缝疲劳破坏机理分析及快速修复技术研究[D].郑州大学,2022.

[20]杜炎霖.基于单目视觉分析的铁路伸缩缝动位移测量方法[D].石家庄铁道大学,2024.

[21]王林林.基于数字图像处理技术的混凝土裂缝特征检测与计算分析[D].山东科技大学,2017.

[22]车紫进.基于机器视觉的混凝土裂缝识别与量化方法研究[D].哈尔滨工业大学,2020.

[23]孙亮,邢建春,谢立强,等.基于自适应阈值Canny算法的裂缝检测方法研究[J].微型机与应用,2017,36(05):35-37+41.

[24]Talab A M A, Huang Z, Xi F, et al. Detection crack in image using Otsu method and multiple filtering in image processing techniques[J]. Optik, 2016, 127(3): 1030-1033.

[25]杨心蕊,许辰扬,郑玉莹,等.基于遗传算法的阈值分割桥梁裂缝检测算法研究[J].广东土木与建筑,2021,28(10):5-9.

[26]Ayenu-Prah A, Attoh-Okine N. Evaluating pavement cracks with bidimensional empirical mode decomposition[J]. EURASIP Journal on Advances in Signal Processing, 2008, 2008: 1-7.

[27]Abdel-Qader I, Abudayyeh O, Kelly M E. Analysis of edge-detection techniques for crack identification in bridges[J]. Journal of computing in civil engineering, 2003, 17(4): 255-263.

[28]张维峰,尹冠生,刘萌,等.数字图像处理技术在桥梁裂纹测量中的应用[J].长安大学学报(自然科学版),2011,31(06):50-53.

[29]徐爱华,皋军.结合图像增强和数学形态学的水泥混凝土路面裂缝识别方法[J].公路,2015,60(10):55-58.

[30]Kheradmandi N, Mehranfar V. A critical review and comparative study on image segmentation-based techniques for pavement crack detection[J]. Construction and Building Materials, 2022, 321: 126162.

[31]Lei Q, Zhong J, Wang C. Joint optimization of crack segmentation with an adaptive dynamic threshold module[J]. IEEE Transactions on Intelligent Transportation Systems, 2024, 25(7): 6902-6916.

[32]隆涛.基于深度学习与图像处理技术的路面裂缝检测方法研究[D].长安大学,2023.

[33]张燚鑫.图像识别的无监督域适应方法研究[D].中国科学技术大学,2022.

[34]Protopapadakis E, Makantasis K, Kopsiaftis G, et al. Crack Identification Via User Feedback, Convolutional Neural Networks and Laser Scanners for Tunnel Infrastructures[C]//VISIGRAPP (4: VISAPP). 2016: 725-734.

[35]梁雪慧,程云泽,张瑞杰,等.基于卷积神经网络的桥梁裂缝识别和测量方法[J].计算机应用,2020,40(04):1056-1061.

[36]张振华,陆金桂.基于改进卷积神经网络的混凝土桥梁裂缝检测[J].计算机仿真,2021,38(11):490-494.

[37]Ding M, Won K. Crack Detection and Location Estimation Using a Convolutional Neural Network[C]//Frontiers of Computer Vision: 27th International Workshop, IW-FCV 2021, Daegu, South Korea, February 22–23, 2021, Revised Selected Papers 27. Springer International Publishing, 2021: 184-188.

[38]Li G, Liu Q, Zhao S, et al. Automatic crack recognition for concrete bridges using a fully convolutional neural network and naive Bayes data fusion based on a visual detection system[J]. Measurement Science and Technology, 2020, 31(7): 075403.

[39]Cha Y J, Choi W, Suh G, et al. Autonomous structural visual inspection using region‐based deep learning for detecting multiple damage types[J]. Computer‐Aided Civil and Infrastructure Engineering, 2018, 33(9): 731-747.

[40]Gao J, Yuan D, Tong Z, et al. Autonomous pavement distress detection using ground penetrating radar and region-based deep learning[J]. Measurement, 2020, 164: 108077.

[41]Wang Y, Song K, Liu J, et al. RENet: Rectangular convolution pyramid and edge enhancement network for salient object detection of pavement cracks[J]. Measurement, 2021, 170: 108698.

[42]李良福,马卫飞,李丽等.基于深度学习的桥梁裂缝检测算法研究[J].自动化学报,2019,45(09):1727-1742.

[43]吴向东,赵健康,刘传奇.基于CNN与CRF的桥梁裂缝检测算法[J].计算机工程与设计,2021,42(01):51-56.

[44]廖延娜,李婉.基于卷积神经网络的桥梁裂缝检测方法[J].计算机工程与设计,2021,42(08):2366-2372.

[45]Alipour M, Harris D K, Miller G R. Robust pixel-level crack detection using deep fully convolutional neural networks[J]. Journal of Computing in Civil Engineering, 2019, 33(6): 04019040.

[46]杨杰文,章光,陈西江,等.基于深度学习的较复杂背景下桥梁裂缝检测[J].铁道科学与工程学报,2020,17(11):2722-2728.

[47]Kang D, Benipal S S, Gopal D L, et al. Hybrid pixel-level concrete crack segmentation and quantification across complex backgrounds using deep learning[J]. Automation in Construction, 2020, 118: 103291.

[48]Peng X, Zhong X, Zhao C, et al. A UAV-based machine vision method for bridge crack recognition and width quantification through hybrid feature learning[J]. Construction and Building Materials, 2021, 299: 123896.

[49]Yang L, Wang B, Zhang R, et al. Analysis on location accuracy for the binocular stereo vision system[J]. IEEE Photonics Journal, 2017, 10(1): 1-16.

[50]Sun X, Jiang Y, Ji Y, et al. Distance measurement system based on binocular stereo vision[C]//IOP Conference Series: Earth and Environmental Science. IOP Publishing, 2019, 252(5): 052051.

[51]Marr D, Poggio T. Cooperative Computation of Stereo Disparity: A cooperative algorithm is derived for extracting disparity information from stereo image pairs[J]. Science, 1976, 194(4262): 283-287.

[52]Ren S, He K, Girshick R, et al. Faster r-cnn: Towards real-time object detection with region proposal networks[J]. Advances in neural information processing systems, 2015, 28.

[53]Sun X, Mei X, Jiao S, et al. Stereo matching with reliable disparity propagation[C]//2011 International Conference on 3D Imaging, Modeling, Processing, Visualization and Transmission. IEEE, 2011: 132-139.

[54]王云峰,吴炜,余小亮,等.基于自适应权重AD-Census变换的双目立体匹配[J].工程科学与技术,2018,50(04):153-160.

[55]闫利,王芮,刘华,等.基于改进代价计算和自适应引导滤波的立体匹配[J].光学学报,2018,38(11):257-267.

[56]郁怀波,胡越黎,徐杰.基于多特征融合与树形结构代价聚合的立体匹配算法[J].上海大学学报(自然科学版),2019,25(01):66-74.

[57]McCulloch W S, Pitts W. A logical calculus of the ideas immanent in nervous activity[J]. The bulletin of mathematical biophysics, 1943, 5: 115-133.

[58]Hinton G E, Osindero S, Teh Y W. A fast learning algorithm for deep belief nets[J]. Neural computation, 2006, 18(7): 1527-1554.

[59]Girshick R, Donahue J, Darrell T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2014: 580-587.

[60]Girshick R. Fast r-cnn[C]//Proceedings of the IEEE international conference on computer vision. 2015: 1440-1448.

[61]Ren S, He K, Girshick R, et al. Faster r-cnn: Towards real-time object detection with region proposal networks[J]. Advances in neural information processing systems, 2015, 28.

[62]He K, Gkioxari G, Dollár P, et al. Mask r-cnn[C]//Proceedings of the IEEE international conference on computer vision. 2017: 2961-2969.