基于数字图像处理技术的平交路口车流量检测-控制理论与控制工程专业论文.docxVIP

摘要摘要 摘要 摘要 随着社会和经济的发展，城市化、汽车化的加快，要求采用现代化的管理 方法来实现交通管理，这样就引起了对智能交通系统(ITS)的研究。ITS能够 改善混乱的交通状况，减少拥堵，提高运输效率和交通安全性。作为ITS的基 本部分，车流量检测系统在ITS中占有很重要的地位。在各种车流量检测系统 中，视频检测方法比其他方法更优越灵活，结合数字图像处理技术，对于交通 安全和交通控制具有非常重要的理论意义和实用价值，是目前研究的热点问题。 本论文详细地研究了车辆目标的视频检测方法和基于投影法的车流量统计 系统。针对车辆目标图像的特点，重点研究了基于背景差分的车辆检测方法； 对图像二值化阈值分割方法中的阈值确定方法进行了研究和比较，选择了效果 较好的0tsu方法；改进了背景获取和背景更新的方法；并对分割后的图像进行 基于形态学的后处理，提高了图像质量；针对车辆的阴影会使图像分割后的目 标区域中含有多余的阴影区域，进而影响检测结果的问题，运用了基于边缘检 测和阴影颜色信息的阴影判定和去除方法：最后设计了一种实时可靠的基于投 影法的车流量统计系统具体算法，并用Visual C++进行编程实现，对采集来的 大量视频图像进行了测试，取得了较好的效果。 关键词：智能交通系统图像处理车辆检测阴影去除流量统计 东北电力大学硕士学位论文Abstract 东北电力大学硕士学位论文 Abstract Modem management methods Rre needed to administer 0n account ofthe society and economic development and the speedup of urbanization and motorization，which begets the research on Intelligent Transport System--ITS．It could develop disorder traffic，reduce traffic jam，improve transportations efficiency and safety．As a basic part of ITS，vehicle detection system plays all important role．In a variety of flow detection systems，vehicle detection based video iS more advanced and flexible thall the other methods，combining digital image processing technology has very important theoretical and practical value for traffic safety and contr01．It becomes a current hot issIle． In this paper,vehicle detection method based video and vehicle flow statistic system based projection are studied in detail．Considering the characters of the moving vehicle pictures，the method of background subtraction is mostly studied to pick up the vehicle．The threshold segmentation in the division method is studied and compared,the Otsu method with better result is chosen．皿e method of background obtainment and update is put forward and studied．And the segmentation image is post—processed based on morphology,which enhances the image quality．To overcome the shadow of moving vehicle will affect the correctness of vehicle recognition and tracking，this paper develops all effective shadow distinguishing and eliminating me