摘要
提出了一种基于并行架构的感兴趣区域提取方法。该方法先采用基于像素色度归类的背景差法定位运动区域,同时利用监视画面的亮度和色度信息提取车牌可能区域;然后将二者的结果进行综合实现车牌区域提取。仿真实验表明,该方法简捷,计算量小,便于用FPGA和DSP快速实现。
A ROI location approach based on parallel framework is presented. The approach firstly uses the method of background subtraction based on pixel hue classification to locate the moving objects. At the same time, the intensity and hue information of the surveillance video are used to locate the possible area of license plate. And then, the results of the above two methods are synthesized to find out the true area of license plate. The experimental results demonstrate that the proposed approach is simple and easy to be realized by DSP and FPGA.
出处
《电视技术》
北大核心
2008年第1期82-84,共3页
Video Engineering
基金
"十一五"国防重点预研项目(102010201)
关键词
感兴趣区域
运动目标
车牌定位
region-of-interest
moving object
license plate location
