Hu和PHOG特征融合的路面标志识别方法  

Road Sign Recognition Method based on Hu and PHOG Feature Fusion

作　　者：张丽艳[1] 张伟[1] 龙美芳 刘鑫 ZHANG Liyan;ZHANG Wei;LONG Meifang;LIU Xin(School of Electrical Information Engineering,Dalian Jiaotong University,Dalian 116028,China)

机构地区：[1]大连交通大学电气信息工程学院,辽宁大连116028

出　　处：《大连交通大学学报》2020年第5期117-120,共4页Journal of Dalian Jiaotong University

基　　金：国家自然科学基金资助项目(51475065)。

摘　　要：对路面交通标识采用支持向量机的方法进行识别,在支持向量机识别交通标志时,需要提取图像的特征进行训练和识别,在Hu不变矩和PHOG特征提取方法的基础上,给出了基于Hu不变矩和低维的PHOG相融合的特征提取方法,该方法在进行路面交通标识识别时,有更好的识别率,仿真结果验证了算法的有效性.Aiming at traffic sign recognition technology in intelligent transportation system(ITS),support vector machine(SVM)is used to recognize road traffic signs.In recognition of traffic signs,image features need to be extracted for training and recognition.On the basis of Hu moment invariant and PHOG feature extraction methods,a fusion method based on Hu moment invariant and low-dimensional PHOG is presented.Feature extraction method has better recognition rate in road traffic identification,and simulation results verify the effectiveness of the algorithm.

关 键 词：HU不变矩 PHOG 特征融合 路面标志识别 

分 类 号：U463.6[机械工程—车辆工程] TP391.41[交通运输工程—载运工具运用工程]