Video based vehicle detection and surveillance technologies are an integral part of Intelligent Transportation System (ITS), due to its non-intrusiveness and capability or capturing global and specific vehicle behavior data. The initial goal of this thesis is to develop an efficient advance warning ITS system for detection of congestion at work zones and special events based on video detection. The goals accomplished by this thesis are: (1) successfully developed the advance warning ITS system using off-the-shelf components and, (2) Develop and evaluate an improved vehicle detection and tracking algorithm. The advance warning ITS system developed includes many off-the-shelf equipments like "Autoscope" (video based vehicle detector), Digital Video Recorders, RF transceivers, high gain Yagi antennas, variable message signs and interface processors. The video based detection system used requires calibration and fine tuning of configuration parameters for accurate results. Therefore, an in-house video based vehicle detection system was developed using the Corner Harris algorithm to eliminate the need of complex calibration and contrasts modifications. The algorithm was implemented using OpenCV library on a Arcom's Olympus Windows XP Embedded development kit running WinXPE operating system. The algorithm performance is for accuracy in vehicle speed and count is evaluated. The performance of the proposed algorithm is equivalent or better to the Autoscope system without any modifications to calibration and lamination adjustments.
展开▼
机译：基于视频的车辆检测和监视技术是智能交通系统（ITS）不可或缺的一部分，因为它具有非侵入性和功能，或者可以捕获全局和特定的车辆行为数据。本文的首要目标是开发一种有效的预警系统，用于基于视频检测的工作区和特殊事件的拥塞检测。本文所要实现的目标是：（1）使用现成的组件成功开发了预警ITS系统，（2）开发和评估了一种改进的车辆检测和跟踪算法。开发的预警ITS系统包括许多现成的设备，例如“ Autoscope”（基于视频的车辆检测器），数字视频录像机，RF收发器，高增益八木天线，可变消息标志和接口处理器。所使用的基于视频的检测系统需要对配置参数进行校准和微调，才能获得准确的结果。因此，使用Corner Harris算法开发了基于内部视频的车辆检测系统，以消除复杂校准和对比度修改的需要。该算法是在运行WinXPE操作系统的Arcom Olympus Windows XP Embedded开发套件上使用OpenCV库实现的。该算法的性能是为了提高车速并评估计数。所提出算法的性能与Autoscope系统相当或更好，无需对校准和层压调整进行任何修改。
展开▼
