文献类型：期刊文献

中文题名：基于主成分特征向量系数的交通标志识别方法研究

英文题名：Study on traffic sign recognition based on principal component eigenvector coefficient

作者：邹柏贤[1];苗军[2];孟斌[1]

第一作者：邹柏贤

机构：[1]北京联合大学应用文理学院;[2]北京信息科技大学计算机学院

第一机构：北京联合大学应用文理学院

年份：2017

卷号：36

期号：24

起止页码：47-50

中文期刊名：微型机与应用

外文期刊名：Microcomputer & Its Applications

基金：国家自然科学基金项目(61650201;41671165);北京市自然科学基金项目(4162058)

语种：中文

中文关键词：交通标志图像;主成分特征向量系数;正确识别率

外文关键词：traffic sign images; principal component eigenvector coefficient; correct recognition rate

摘要：对交通标志的识别研究一直是模式识别领域的研究热点。提出一种利用主成分特征向量系数和最近邻分类识别交通标志的方法,经验证取得较好的识别效果;同时,还研究探讨了交通标志图像的分辨率大小、主成分特征个数对正确识别率的影响。该方法的特点是交通标志图像来自真实环境,减小了计算量。
The study on traffic sign recognition is always a research hotspot in the field of pattern recognition. A new method to recognize traffic signs using principal component eigenvector coefficient and the nearest neighbor classification was proposed. It was proved that the method has better recognition effect. The proposed method was simple and proved to be effective with traffic sign image in the real environment. Through experiments, the relationship between the resolution and the recognition rate of traffic sign images was discussed, and the relationship between the number of the principal components and the recognition rate was analyzed too. Reached a conclusion, when the resolution of the traffic sign image was 64 × 64, the recognition rate reached the maximum using one principal component feature.