基于模糊形状判别的鲁棒交通标志检测算法  

Robust Traffic Sign Detection Using Fuzzy Shape Recognizer

作　　者：丁淑艳[1] 李伦波[1] 潘海朗[1] 

机构地区：[1]南京理工大学电子工程与光电技术学院,南京210094

出　　处：《机电一体化》2012年第12期13-18,48,共7页Mechatronics

基　　金：国家自然科学基金(61075031)

摘　　要：针对交通标志存在尺度、旋转、倾斜、表面被污损或部分被遮挡等退化情况,提出一种基于模糊形状判别的鲁棒交通标志检测算法。该算法在HSV彩色空间进行颜色分割的基础上,提取目标区域的对称局部特征,根据设计的模糊形状判别算法来判定目标区域的形状,进而检测出交通标志。采用该检测算法对不同天气情况、不同道路场景下的3 000多幅自然场景图像进行交通标志检测实验,结果表明该检测算法不仅具有较高的检测率,而且具有良好的鲁棒性。A novel traffic sign detection algorithm based on fuzzy shape recognizer is proposed for the detection of traffic signs in natural environments, which usually suffer from various forms of image degradation, such as scaling, rotation, tilt, damage, and partial occlusion. Every RGB image is converted into HSV color space, and segmented by the hue and saturation thresholds. The local features of the regions of interest （ROI） are extracted using a symmetrical detector, and the shape of ROI is determined by a fuzzy shape recognizer. More than 3 000 road images were collected under different weather conditions and different road scenes, and used for testing this approach. Experimental results show that the proposed algorithm not only obtains high detection accuracy, but also has good robust performance.

关 键 词：辅助驾驶系统 交通标志检测 HSV彩色空间 模糊形状判别 

分 类 号：U463.6[机械工程—车辆工程] U495[交通运输工程—载运工具运用工程]