基于视频的电子警察系统性能分析与研究

本文详细分析了目前基于视频的电子警察抓拍系统的性能,指出了电子警察抓拍系统中存在的不足:系统启动时,需要长期的环境适应,即机器人智能训练;对正常的左转、正常逆行车辆和违章车辆难以区分;对于车辆稀少的路口性能表现还可以满足要求,但对于车辆和人行比较密集的城市交通要道,其表现往往不如人意,存在闯红灯误拍率高等缺点,本文针对以上问题,提出了自己的解决方案,据此设计了新的系统,实验表明:新系统的稳定性和抓拍的准确率都有了明显的改善。

Analysis and Research on Performance of E-Police System Based on Vidoe Testing Technology

This paper analyses the performance of the present E-Police systems based on video testing technology in detail,and it shows the shortages existing in the present E-Police systems: when they were used in the first time,they needed a long time to adapt the circumstance,intelligent training like a robot. These systems are very difficult to classify the normal vehicles which turn left or run backward. They had good behavior when they were used in light traffic crossings,however,when under heavy traffic conditions,they couldn't meet the basic needs and make a lot of invalidated pictures. In this paper,we give the new solutions to these problems and design the new system,By running under trial,the robust and the ratio of accurately snatching are apparently improved.