动态称重系统作为交通管理的一项先进技术，旨在在不妨碍交通流动的情况下，对行驶中的车辆进行重量测量。该系统在货运车辆超载监控、基于重量的收费以及评估道路与桥梁的结构健康状况等方面发挥着至关重要的作用。然而，动态称重系统在实际应用中由于路面交通环境复杂，虽然在该系统正式投入使用之前进行了精准的参数标定，但仍然存在部分车辆不能够精确称重的现象。由于不同的驾驶员有不同的驾驶行为，当不同类型和尺寸的车辆通过动态称重区域时，车辆的运动状态变化是影响称重精度的主要原因。本文通过对影响称重精度的多个因素进行深入分析，明确了导致称重误差的主要原因，识别和分析车辆尺寸和运动状态对称重准确度的具体影响，特别是对于长度约为12.5米的车辆在不同运动状态下的称重精度问题，并针对这些问题提出有效的误差校正策略，并提出相应的误差补偿方法。研究的主要内容有：

（1）基于车辆信息的动态称重系统精度补偿框架。首先详细分析了车辆驶过动态称重传感器的称重过程，然后明确了导致系统称重误差的多种潜在因素。在此基础上，为了提高动态称重系统的精度，提出了一个基于车辆信息的精度补偿方法。最后，对所提出的动态称重系统精度补偿方法进行了概括性描述，详细说明了每一部分的功能及其在整个补偿流程中的作用。

（2）动态称重区域车辆尺寸测量方法。针对现有尺寸测量方法成本高、算法复杂不能满足实时性要求的问题，提出了基于车辆轮廓的尺寸预测方法。首先使用优化后的YOLO V5s检测出动态称重区域车辆并识别出车辆轴数，以便后续分析尺寸与车辆类型、动态称重结果之间的关系。然后利用CEDN网络提取出目标车辆的轮廓，建立车辆尺寸与轮廓数据集，最后使用改进的DenseNet网络预测出车辆的长宽高。

（3）动态称重区域车辆的运动状态信息提取方法。针对现行多目标跟踪方法在对车辆遮挡的高度敏感性和在多目标跟踪过程中频繁的车辆ID变换问题，首先提出一种改进的SSD车辆检测模型，以减少车辆遮挡引起的检测遗漏问题。其次，为了获得更加稳定的轨迹和速度信息，在DeepSORT算法的基础上引入了高斯平滑插值（GSI）方法。最后，通过一系列实验，对所提出的改进模型在动态称重区域车辆跟踪任务中的性能进行了验证。

（4）融合动态称重数据进行车辆尺寸和运动状态变化对称重精度的影响分析与补偿实验验证。首先，构建了一个包含动态称重区域车辆长宽高及速度波动信息的数据集，接着分析了影响动态称重准确性的关键因素。随后，针对长度在12.5米左右的车辆在不同速度波动下对称重精度的影响实施了补偿策略，以验证本文提出的基于车辆信息的补偿方法的可行性。

Weigh-in-motin Systems (WIMs) represent an advanced technology in traffic management, designed to measure the weight of moving vehicles without impeding traffic flow. These systems play a pivotal role in the monitoring of overloaded freight vehicles, implementation of weight-based tolling, and assessment of structural health for roads and bridges. However, despite precise calibration prior to their deployment, WIMs often face challenges in achieving accurate vehicle weighing in complex road traffic environments. This discrepancy in performance is primarily due to the variations in vehicle motion states caused by different driving behaviors, especially when vehicles of diverse types and sizes traverse the weighing area. This thesis conducts a thorough analysis of the factors affecting weighing accuracy, identifies the main causes of weighing error, and analyzes the specific impacts of vehicle size and motion state on the accuracy of the weighing process. The study pays particular attention to the weighing precision issues for vehicles approximately 12.5 meters in length under various motion states. In response to these issues, the thesis proposes effective error correction strategies and corresponding compensation methods. The main contents of the research include:

(1) A vehicle information-based precision compensation framework for WIMs.Firstly, the thesis comprehensively analyzes the process of vehicles passing through the dynamic weighing sensor, and then determines various potential factors that lead to the system weighing error. On this basis, in order to improve the accuracy of WIMs, a precision compensation method based on vehicle information is proposed. Finally, the proposed WIMs precision compensation method is summarized, and the function of each component and its role in the whole compensation process are explained in detail.

(2) Vehicle size measurement method in the weigh-in-motion region. In response to the high costs and complexity of existing size measurement methods, which fail to meet real-time requirements, a size prediction method based on vehicle contours is proposed. Initially, the optimized YOLO V5s algorithm is employed to detect vehicles within the weigh-in-motion region and identify the number of axles, facilitating subsequent analysis of the relationship between size and vehicle type, as well as dynamic weighing results. Subsequently, the CEDN network is utilized to extract the contours of the target vehicle, and a dataset correlating vehicle size with contour data is established. Finally, an improved DenseNet network is used to predict the dimensions (length, width, and height) of the vehicle.

(3) Method for extracting motion state information of vehicles in weigh-in-motion region. Addressing the high sensitivity to vehicle occlusions and the frequent vehicle ID switches during multi-object tracking in current methodologies, an enhanced SSD (Single Shot MultiBox Detector) vehicle detection model is initially proposed to mitigate detection omissions caused by vehicle occlusions. Subsequently, to obtain more stable trajectory and velocity information, a Gaussian Smooth Interpolation (GSI) method is introduced on top of the DeepSORT algorithm. Finally, a series of experiments are conducted to validate the performance of the proposed improved model in vehicle tracking tasks in weigh-in-motion region.

(4) Integration of dynamic weighing data for the analysis and experimental validation of the impact of vehicle size and motion state variations on weighing accuracy, and compensation thereof. Firstly, a data set containing the information of vehicle length, width, height and speed fluctuation in the dynamic weighing area is constructed. Subsequently, the key factors affecting the accuracy of weigh-in-motion are analyzed. Following this, a compensation strategy is implemented for vehicles approximately 12.5 meters in length under different velocity fluctuations to assess the impact on weighing accuracy, thereby validating the feasibility of the vehicle information-based compensation method proposed in this thesis.

[1] 2022年交通运输行业发展统计公报[J]. 中国水运, 2023(7): 29-33.

[2] 贺玲玲. 车辆超限超载对公路桥梁的影响[J]. 山西交通科技, 2020(4): 151-153.

[3] 付涛. 车辆超限超载对桥梁的影响与应对措施[J]. 工程技术研究, 2021, 6(11): 159-160.

[4] 超限运输车辆行驶公路管理规定[J]. 中华人民共和国国务院公报, 2016(29): 29-36.

[5] 交通运输部关于修改《超限运输车辆行驶公路管理规定》的决定[J]. 中华人民共和国国务院公报, 2021(29): 87.

[6] 关于在全国开展车辆超限超载治理工作的实施方案[J]. 交通标准化, 2004(9): 12-16.

[7] Haugen T, Levy J R, Aakre E, et al. Weigh-in-motion equipment experiences and challenges[J]. Transportation Research Procedia, 2016, 14: 1423-1432.

[8] 赵培杰. 基于压电石英传感器的高速动态称重系统设计[D]. 中北大学, 2018.

[9] 李扬. 国外汽车超载治理技术的发展及应用[J]. 交通标准化, 2004(9): 80-82.

[10] Jacob B, O’Brien E J. European specification on weigh-in-motion of road vehicles (COST323)[C]//Proceedings of second European conference on weigh-in-motion of road vehicles, Held Lisbon, Portugal. 1998: 14-16.

[11] 王荣旭. 基于压电传感器的车辆动态称重系统开发[D]. 山东大学, 2021.

[12] 张文斌. 公路车辆动态荷载测量及车型分类技术的研究[D]. 哈尔滨工业大学, 2011.

[13] 武奇生, 王丹, 陈圆媛, 等. 基于ETC的车辆动态称重系统设计[J]. 公路交通科技(应用技术版), 2009, 5(10): 187-189+194.

[14] 孙婧. 信念引领“自由”——北京万集科技有限责任公司自由流产品成功应用于武汉路桥收费[J]. 中国交通信息化, 2011(10): 134-136.

[15] 刘劲松, 叶程. 基于窄条传感器的高精度动态称重系统设计和实现[J]. 中国交通信息化, 2016(3): 99-100+115.

[16] 王平, 邓永强, 陈忠元, 等. 高速公路收费站入口治超系统方案[J]. 中国交通信息化, 2019(3): 97-100.

[17] 张惠芳, 张海宁. 动态车辆称重系统的算法研究[J]. 国外电子测量技术, 2017, 36(7): 52-54+61.

[18] 梁杰. 基于DSP的动态电子称重系统的设计[D]. 哈尔滨理工大学, 2017.

[19] Zhang R, Lv W hong, Guo Y jing. A Vehicle Weigh-in-Motion system based on hopfield neural network adaptive filter[C]//2010 International Conference on Communications and Mobile Computing. Shenzhen, China: IEEE, 2010: 123-127.

[20] 沈小倩. 新型自适应滤波处理方法在动态称重中的应用[J]. 中国计量大学学报, 2021, 32(2): 184-188.

[21] Xu S, Chen X, Fu Y, et al. Research on weigh-in-motion algorithm of vehicles based on BSO-BP[J]. Sensors, 2022, 22(6): 2109.

[22] 陈笑颖. 基于神经网络车辆动态称重技术研究[D]. 西安工业大学, 2023.

[23] Qin T, Lin M, Cao M, et al. Effects of Sensor Location on Dynamic Load Estimation in Weigh-in-Motion System[J]. Sensors, 2018, 18(9): 3044.

[24] 赵千. 基于多传感器的路面动态称重系统研究[D]. 北京科技大学, 2020.

[25] Klein E, Stanczyk D, Ieng S S. Improvement of weigh-in-motion accuracy by taking into account vehicle lateral position[C]//Proc., 6th Int. Conf. on Weigh-in-Motion, Jacob, B., ed., John Wiley & Sons, Dallas. 2012: 206-213.

[26] Burnos P, Gajda J, Sroka R, et al. High Accuracy Weigh-In-Motion Systems for Direct Enforcement[J]. Sensors, 2021, 21(23): 8046.

[27] Gajda J, Sroka R, Stencel M, et al. Design and accuracy assessment of the multi-sensor weigh-in-motion system[C]//2015 IEEE International Instrumentation and Measurement Technology Conference (I2MTC) Proceedings. 2015: 1036-1041.

[28] Burnos P, Gajda J. Thermal Property Analysis of Axle Load Sensors for Weighing Vehicles in Weigh-in-Motion System[J]. Sensors, 2016, 16(12): 2143.

[29] 高彤彤. 基于压电薄膜的车辆动态称重系统的研究与实现[D]. 山西大学, 2019.

[30] 徐志玲, 沈裴裴, 厉志飞. 车辆动态称重的速度补偿研究[J]. 计量学报, 2015, 36(6): 599-602.

[31] Ryguła A, Maczyński A, Brzozowski K, et al. Influence of Trajectory and Dynamics of Vehicle Motion on Signal Patterns in the WIM System[J]. Sensors, 2021, 21(23): 7895.

[32] 何海浪. 压电式车辆动态称重传感器的设计与研究[D]. 杭州电子科技大学, 2016.

[33] 王中立, 李丽宏. 基于石英传感器的动态称重数据处理算法研究[J]. 传感技术学报, 2017, 30(2): 236-241.

[34] Wang Z, Wang Q, Hou J. The Research of Weigh-in-Motion Sensor Layout Based on Data Fusion[C]//2009 Sixth International Conference on Fuzzy Systems and Knowledge Discovery: Vol. 2. 2009: 127-131.

[35] 罗雁, 苏清祖, 陈晓东, 等. 汽车行驶称重系统的算法研究与精度分析[J]. 汽车工程, 2005(4): 495-497+422.

[36] 杜磊. 基于车路耦合系统的路面动力分析及疲劳寿命研究[D]. 重庆交通大学, 2018.

[37] 李涛, 卢海波, 林泛业, 等. 重心高度对车辆动力学性能的影响[J]. 汽车实用技术, 2019(3): 110-112.

[38] 唐俊义, 冯麟, 周志祥, 等. 基于单目视觉测量技术的车辆尺寸识别方法[J]. 公路交通科技, 2023, 40(3): 228-236.

[39] Ratajczak R, Grajek T, Wegner K, et al. Vehicle dimensions estimation scheme using AAM on stereoscopic video[C]//2013 10th IEEE International Conference on Advanced Video and Signal Based Surveillance. Krakow, Poland: IEEE, 2013: 478-482.

[40] 孔月萍, 赵天悦, 王佳婧. 动态无接触的货车尺寸测量方法[J]. 激光与光电子学进展, 2020, 57(21): 173-178.

[41] Li S, Han L, Dong P, et al. Algorithm for measuring the outer contour dimension of trucks using UAV binocular stereo vision[J]. Sustainability, 2022, 14(22): 14978.

[42] 梁春疆, 段发阶, 杨毅, 等. 车辆外廓尺寸计算机视觉动态测量[J]. 光电工程, 2016, 43(1): 42-48+54.

[43] 张斌. 基于逆投影面图像匹配的车辆外廓测量技术研究[D]. 长安大学, 2019.

[44] Dai F, Park M W, Sandidge M, et al. A vision-based method for on-road truck height measurement in proactive prevention of collision with overpasses and tunnels[J]. Automation in Construction, 2015, 50: 29-39.

[45] Lu L, Dai F. Automated visual surveying of vehicle heights to help measure the risk of overheight collisions using deep learning and view geometry[J]. Computer-Aided Civil and Infrastructure Engineering, 2023, 38(2): 194-210.

[46] Nguyen B, Brilakis I, Vela P A. Optimized parameters for over-height vehicle detection under variable weather conditions[J]. Journal of Computing in Civil Engineering, 2017, 31(5): 04017039.

[47] Wang Z, Zhan J, Duan C, et al. A review of vehicle detection techniques for intelligent vehicles[J]. IEEE Transactions on Neural Networks and Learning Systems, 2023, 34(8): 3811-3831.

[48] Dong X, Yan S, Duan C. A lightweight vehicles detection network model based on YOLOv5[J]. Engineering Applications of Artificial Intelligence, 2022, 113: 104914.

[49] Woo S, Park J, Lee J Y, et al. CBAM: convolutional block attention module[M]. arXiv, 2018.

[50] 简柯青. 目标物体的轮廓识别关键技术研究[D]. 电子科技大学, 2022.

[51] Jian K, Gui S. Object and contour detection with an architecture-fusion network[C]//2021 IEEE 33rd International Conference on Tools with Artificial Intelligence (ICTAI). Washington, DC, USA: IEEE, 2021: 910-914.

[52] Everingham M, Van Gool L, Williams C K I, et al. The pascal visual object classes (VOC) challenge[J]. International Journal of Computer Vision, 2010, 88(2): 303-338.

[53] Huang G, Liu Z, Van Der Maaten L, et al. Densely connected convolutional networks[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI: IEEE, 2017: 2261-2269.

[54] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, NV, USA: IEEE, 2016: 770-778.

[55] Howard A G, Zhu M, Chen B, et al. MobileNets: efficient convolutional neural networks for mobile vision applications[EB]//arXiv.org. (2017-04-17).

[56] Bewley A, Ge Z, Ott L, et al. Simple online and realtime tracking[C]//2016 IEEE International Conference on Image Processing (ICIP). 2016: 3464-3468.

[57] Wojke N, Bewley A, Paulus D. Simple online and realtime tracking with a deep association metric[C]//2017 IEEE International Conference on Image Processing (ICIP). 2017: 3645-3649.

[58] 金立生, 华强, 郭柏苍, 等. 基于优化DeepSort的前方车辆多目标跟踪[J]. 浙江大学学报(工学版), 2021, 55(6): 1056-1064.

[59] Gao B. Research on two-way detection of YOLO V5s+Deep Sort road vehicles based on attention mechanism[J]. Journal of Physics: Conference Series, 2022, 2303(1): 012057.

[60] Zhang K, Wu F, Sun H, et al. Monocular vehicle speed detection based on improved YOLOX and DeepSORT[J]. Neural Computing and Applications, 2023.

[61] Chen W, Qiao Y, Li Y. Inception-SSD: An improved single shot detector for vehicle detection[J]. Journal of Ambient Intelligence and Humanized Computing, 2022, 13(11): 5047-5053.

[62] Cao J, Song C, Song S, et al. Front vehicle detection algorithm for smart car based on improved SSD model[J]. Sensors, 2020, 20(16): 4646.

[63] 曹孟新. 基于改进SSD算法的车辆目标检测方法研究与系统实现[D]. 安徽工程大学, 2023.

[64] Du Y, Zhao Z, Song Y, et al. StrongSORT: make DeepSORT great again[M]. arXiv, 2023.

[65] Guo F, Xu Y. Vehicle analysis system based on DeepSORT and YOLOv5[C]//2022 3rd International Conference on Computer Vision, Image and Deep Learning & International Conference on Computer Engineering and Applications (CVIDL & ICCEA). Changchun, China: IEEE, 2022: 175-179.