摘要
鉴于红绿灯在城市交通中占有的重要地位,通过图像处理设计一个系统来辅助驾驶员判别红绿灯。首先通过图像采集及图像预处理的方法获得源图像,然后再设立红绿灯主要特征训练集,用边缘跟踪的方法框选出红绿灯类似目标,对类似目标进行特征提取,再与标准模板做对比,确定为目标后标出各个灯的位置。最后通过图像坐标读取各个灯的颜色分量,比较灯的各色分量,从而确定每一个灯的颜色。识别完成后将识别结果通过数字显示或声音提示的方式输出以提醒驾驶员注意红绿灯的变化。
Due to traffic lights occupies an important position in the city transportation, design a system to assist the drivers note the traffic lights＇ changing through image processing in this paper.First of all, get the source iraages with the method of image acquisition and image preprocessing, then set up the main characteristic training of traffic lights. Choose the goals which is similar with traffic lights with the method of the edge tracing. After extracting the features of similar target and contrasting with the standard template, determine the target and each lamp＇s coordinates. Finally, read out each light＇s color component through the image coordinates and judge each light＇s color with the way of comparing evely light＇s color component. After recognition,identify the results through the way of digital display or voice prompt to remind drivers pay attention to the traffic lights changing.
出处
《微型机与应用》
2013年第4期88-90,93,共4页
Microcomputer & Its Applications
关键词
红绿灯
模式识别
图像处理
智能交通
traffic lights
pattern recognition
image processing
intelligent transportation
作者简介
童刚，男，1962年生，教授，主要研究方向：城市智能交通．
朱鸿博，男，1987年生，硕士研究生，主要研究方向：图像处理与智能交通。
