期刊论文

Traffic signs recognition based on geometry invariant moments

中文

武汉轻工大学学报

2095-7386

2011

30

4

58-61

10.3969/j.issn.1009-4881.2011.04.015

湖北省教育厅科研计划项目（2002D01001）；

本校为第一机构

以警告交通标志图像为研究对象,首先采用Hu不变矩来提取交通标志图像特征,然后利用最小矩距离法和相关系数法对交通标志进行识别。实验结果表明,该方法具有很好的识别能力,对旋转、缩放后的图象可以达到较高的识别率,基于Hu不变矩的几何特征提取算法在图像识别中的应用是非常有效的。

Rresearch object of this paper is warning traffic signs images.Firstly,features of traffic signs images are extracted using Hu＇s invariant moments.Then,traffic signs are recognised using method of minimal distance and method of correlative coefficient.Experiment results prove that the method has effective recognition ability.The recognition rate may achieve 100% for rotated and zoomed images.Algorithm of extracting geometrical characteristics based on Hu＇s...