基于云计算视觉技术的肇事车辆定位系统设计  

Hit-and-Run Vehicle Positioning System Design Based on Cloud Computing Vision Technology

作　　者：杨志[1] 胡文红[1] 

机构地区：[1]浙江水利水电专科学校,杭州310018

出　　处：《计算机测量与控制》2013年第3期770-772,共3页Computer Measurement &Control

摘　　要：针对车辆在肇事逃逸过程中,运动车辆图像所在背景变换较为快速,图像中有大量的相似性特征干扰,使得车辆可定位的特征发生变化;传统的车辆定位系统缺乏应对这种背景变化和像素干扰的能力,系统的计算量较大造成延时,造成定位效果不准确;为此提出一种基于云计算视觉技术的肇事车辆定位系统设计方法;通过视觉采集系统对逃逸的可疑车辆进行图像的实时采集,运用信息传递模块发送特定信息到终端,运用云计算技术,完成海量信息处理和车辆定位;实验证明,这种系统设计方法能够大幅提高肇事逃逸车辆图片的关键帧定位结果的准确性,定位准确性达到65%。For vehicles in escapes process, moving vehicle in the background image a more rapid transformation, images, there are a lot of similarities interference characteristics, which allows the positioning of the characteristic change. The traditional vehicle positioning system lack of response to the background changes and the ability of pixels interference, the amount of calculation system caused large time delay, cause orientation effect is not accurate. A design method of hit and--run vehicle positioning system is proposed based on a cloud computing vision technology. Through the visual collection system to the suspicious vehicle of the image of real time data collection, the use of informa- tion transmission module send specific information to the terminal, using cloud computing technology, complete the mass information pro- cessing and vehicle positioning. Experiments show that the system design method can greatly improve the spin off vehicle picture key frames the accuracy of the locating results, positioning accuracy to 65 %.

关 键 词：肇事逃逸 云计算 车辆定位 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]