导　　师： 权义宁

学科专业： 081201

授予学位： 硕士

作　　者： ;

机构地区： 西安电子科技大学

摘　　要： 随着高速公路网建设的日臻完善,利用较为完备的视频监控系统对路面状况进行实时监控成为提高道路运营质量的一种重要手段。基于视频图像的目标检测和数据采集分析为更大程度的发挥监控系统的作用提供了必要的技术支持,其相对于传统监控手段具有智能化程度高、信息多元化、投资回报率高等优点。 本文基于数字图像处理和计算机视觉技术,对视频图像中的目标运动模式进行了分析和研究,通过对运动目标检测、目标跟踪、交通流参数提取的算法研究,完成了对视频图像中热点区域目标的监控,并对目标的运动行为进行了过滤和分类,完成了交通流参数的统计分析。利用本文算法,通过在实验路段上对交通流参数进行检测,结果表明具有较高的检测精度,能够为交通运营管理提供必要的技术支持,达到了提高路网运营效率的目的。 本文的研究成果体现在,采用分割技术将道路运行车道作为监控重点区域,提高了检测效率,抑制了噪声干扰,降低了系统误报率和错报率；采用基于特征区域的卡尔曼滤波算法对目标进行跟踪,实现了复杂背景下运动目标的快速跟踪,提高了跟踪效率,有较强的鲁棒性。在此基础之上开发了交通流检测系统,实现了视频监控系统的正常、有效运转,并在实际工程中得到应用,系统运行稳定、安全、可靠,具有较高的实际应用价值。 Along with improving the construction of the highway network, the use of a more complete video surveillance system for real-time monitoring of road conditions as to improve the quality of the road is an important means of operation. Target detection based on video image analysis and data collection play a greater role in monitoring system to provide the necessary technical support relative to traditional means of monitoring a high degree of intelligence, information diversity, the advantages of high return on investment. Based on digital image processing and computer vision technology, in the video image of target motion model for the analysis and research, through the moving target detection, target tracking and traffic parameter extraction algorithm, completed the target hot spot areas in video surveillance, and target motion behavior of the filtering and classification, completed the statistical analysis of traffic flow parameters. Using this algorithm, the experimental parameters of traffic flow on the section tested, the results show that high detection accuracy, for traffic operation and management to provide the necessary technical support, achieve the purpose of improving the road network operational efficiency. The results of this study, using segmentation of road running lane as monitoring key areas to improve the detection efficiency, suppress noise and reduce the system false alarm rate and the rate of false positives; using Kalman filtering algorithm based on feature area to track the target, enabling fast tracking of moving targets in complex background, increased tracking efficiency, strong robustness. Traffic flow detection system based on this development, video surveillance system to achieve a normal and effective operation, and has been applied in practical engineering, the system is stable, safe, reliable and has high practical value.

关 键 词： 交通流参数 视频检测 图像处理 目标检测 目标跟踪

分 类 号： [TP391.41 TP274]

领　　域： [自动化与计算机技术] [自动化与计算机技术] [自动化与计算机技术] [自动化与计算机技术]