基于目标跟踪的车流量统计系统研究

发布时间：2019-05-14 23:19

【摘要】：基于视频图像的车流量统计系统是智能交通系统的重要研究课题之一,它利用图像处理和人工智能技术对摄录的车流视频进行分析处理,获取在某个时段中道路上通过的车辆数目,为智能交通系统提供后续处理的基本数据,实现道路智能化调度,提高路面资源利用率。基于目标跟踪的车流量统计方法在实际应用中,由于受到光照的影响经常会出现阴影问题,当阴影区域面积较大和车辆间行驶距离较近时,会造成图像中车辆发生粘连,从而影响车流量统计系统的准确率。因此,车辆阴影消除算法、粘连车辆分割算法是车流量统计系统的关键技术。本文在研究车流量统计过程中常用的阴影消除和粘连分割算法的基础上,实现了一个基于目标跟踪的车流量统计系统。本文主要研究工作如下:1.针对车辆边缘的遮挡车辆曲线分割算法在分割过程中准确率低的问题,改进了一种基于凹性分析的粘连车辆分割算法。仿真实验结果表明,本文的算法与车辆边缘的遮挡车辆曲线分割算法相比,在分割车辆时具有较高的准确率,从而可以有效地提高车流量统计系统的准确性。2.对常用的阴影消除算法(HSV颜色特征算法、梯度特征算法)进行了分析研究。针对上述两种算法各自消除率低的问题,提出了一种基于HSV颜色特征和梯度特征融合的阴影消除算法。仿真实验结果证明,本文所提出的算法具有较高的阴影消除率,在一定程度上提高了车辆检测的鲁棒性。3.在Visual Studio 2010集成开发环境下,采用MFC应用程序框架及Open CV计算机视觉库,实现了基于目标跟踪的车流量统计系统。实验结果表明,本文所设计的系统具有较好的实时性,可以在一定程度上克服光照的影响。
[Abstract]:The traffic flow statistics system based on video image is one of the important research topics of intelligent transportation system. It uses image processing and artificial intelligence technology to analyze and process the recorded traffic flow video. The number of vehicles passing through the road in a certain period of time is obtained, the basic data of follow-up processing are provided for the intelligent transportation system, the intelligent dispatching of the road is realized, and the utilization rate of pavement resources is improved. In practical application, the traffic flow statistics method based on target tracking often has shadow problem because of the influence of illumination. when the shadow area is large and the driving distance between vehicles is close, the vehicle adhesion will occur in the image. Thus, the accuracy of traffic flow statistics system is affected. Therefore, vehicle shadow elimination algorithm and adhesion vehicle segmentation algorithm are the key technologies of traffic flow statistics system. Based on the study of shadow elimination and adhesion segmentation algorithms commonly used in traffic flow statistics, a traffic flow statistics system based on target tracking is implemented in this paper. The main research work of this paper is as follows: 1. In order to solve the problem of low accuracy of vehicle curve segmentation algorithm for vehicle edge occlusion, an adhesive vehicle segmentation algorithm based on concave analysis is improved. The simulation results show that compared with the vehicle edge occlusion vehicle curve segmentation algorithm, the proposed algorithm has higher accuracy in vehicle segmentation, which can effectively improve the accuracy of the traffic flow statistics system. 2. The common shadow elimination algorithms (HSV color feature algorithm, gradient feature algorithm) are analyzed and studied. In order to solve the problem of low elimination rate of the above two algorithms, a shadow elimination algorithm based on the fusion of HSV color features and gradient features is proposed. The simulation results show that the algorithm proposed in this paper has a high shadow elimination rate and improves the robustness of vehicle detection to a certain extent. Based on Visual Studio 2010 integrated development environment, a traffic flow statistics system based on target tracking is implemented by using MFC application framework and Open CV computer visual library. The experimental results show that the system designed in this paper has good real-time performance and can overcome the influence of light to a certain extent.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491;TP391.41

【参考文献】