基于视频图像像素点阵的车辆检测技术  

Vehicle Examination Method that Based on Video Pixel-matrix

作　　者：王祺[1] 游志胜[1] 覃树建[1] 

机构地区：[1]四川大学计算机学院,四川成都610064

出　　处：《计算机应用研究》2006年第8期183-185,共3页Application Research of Computers

摘　　要：介绍一种基于视频图像像素点阵的车辆检测技术。该方法找出车辆图像中因与背景帧不同而触发的像素点,通过分析这些触发像素点所组成的像素点阵进行对车辆的自动识别和追踪。将每辆车作为一个检测对象,求出每个检测对象的触发像素点阵中存在连通性的最大像素区域,并将这个区域作为该检测对象识别和追踪的唯一标准。该技术具有适应性好、计算量少、识别结果准确等优点。This text introduce a vehicle examination method that base on the video frequency pictures that grasp to clap. This method find out the triggered pixels because of different from the background. Through analytical these pixel-matrixes that constitute of triggered pixels to identify and track the vehicle automatically. And be the each vehicle as an examination object, find out a biggest pixel district in pixel-matrix of each examination object, and this district is the unique standard of identifying and tracking a vehicle object. That system has many advantages, such as good adaptability, little calculation, identifying result is accurate etc.

关 键 词：视频图像 像素点阵 触发稳定区 车辆对象 车辆检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]