论文部分内容阅读
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。
This paper analyzes the performance of video-based electronic police capture system in detail and points out the shortcomings of the electronic police capture system: when the system is started, it needs long-term adaptation to the environment, that is, robot intelligent training; for normal left turn, normal reverse vehicle And the illegitimate vehicles are indistinguishable; for the sparseness of vehicles, the intersection performance can meet the requirements, but for the vehicles and the densely populated urban traffic arteries, the performance is often unsatisfactory, there is the shortcomings of red light running high error rate. In view of the above problems, Proposed its own solution, and accordingly designed a new system, experiments show that: the stability of the new system and capture the accuracy of the rate has been significantly improved.