基于HSV颜色特征的车辆阴影消除方法  

The Elimination Methods of Vehicle Shadow Based on HSV Color Features

作　　者：许宏科[1] 侯晓青[1] 秦严严[1] 

机构地区：[1]长安大学电子与控制工程学院,陕西西安710064

出　　处：《徐州工程学院学报（自然科学版）》2016年第2期5-8,共4页Journal of Xuzhou Institute of Technology(Natural Sciences Edition)

基　　金：陕西省交通运输厅科研项目(12-26K);国家山区公路工程技术研究中心开放基金项目(gsgzj-2011-08)

摘　　要：简单使用HSV颜色特征进行车辆阴影消除时,在车辆和道路环境背景颜色信息较相似的情况下,易将车辆本身的若干像素错误地当作阴影消除.鉴于此,首先对当前图像和背景图像逐个像素地进行对数运算,得到对应的Log域图像,并对该2幅Log域图像进行差分操作,从而起到弱化车辆阴影的目的.然后,在阴影得到弱化的基础上,使用HSV颜色特征检测出车辆候选区域中的阴影区域,并给予剔除,最终达到阴影消除的目的.仿真实验效果良好,综合反映阴影消除方法性能的评价指标均值约为90.10%.Vehicle shadow elimination can result in error elimination of a number of pixels belonging to the vehicle itself using HSV color features if vehicle color information is similar to that of road environment background. In view of this, firstly, the current image and the background image were calculated logarith- mic for every pixel, which can get the corresponding log domain images. And image difference was made on the two log domain images for softening the vehicle shadow. Secondly, shadow areas were detected from vehicle candidate areas using HSV color features and were eliminated for purpose of shadow elimina- tion. The simulation result is good, and the evaluation index that reflects the performance of shadow elimi- nation methods is about 90.10%.

关 键 词：车辆阴影消除 HSV颜色模型 对数运算 阴影评价 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]