基于PCA和灰度直方图特征融合的交通标志的分类研究

发布时间：2018-04-23 13:02

  本文选题：交通标志 + 识别　； 参考：《公路》2017年04期

【摘要】：为了提高交通标志分类问题的正确率,提取有效的特征值才可以获得更高的分类正确率。校核分析交通标志图像特点,在分类研究的背景下提出了特征融合的思路,在主成分分析(PCA)降低维度的基础上,提取灰度直方图的特征,将PCA提取的特征和灰度直方图特征融合,并且将融合数据作为分类的输入特征,通过交通标志数据库进行实验分析,多次改变要降低的维度,然后融合灰度直方图特征进行分类,用MATLAB和GUI工具进行仿真,实例验证结果表明,得出的正确率明显提高,在交通标志的分类中效果显著。
[Abstract]:In order to improve the correct rate of traffic sign classification problems and extract effective eigenvalues, a higher classification accuracy can be obtained. The characteristics of traffic sign images are analyzed and the idea of feature fusion is put forward in the background of classification research. On the basis of the dimension reduction of principal component analysis (PCA), the features of gray histogram are extracted and PCA is extracted. The characteristics and gray histogram feature fusion, and use the fusion data as the input characteristics of the classification, through the traffic sign database experiment analysis, many times change to reduce the dimension, then fusion gray histogram characteristics to classify, using MATLAB and GUI tools to simulate. Example verification results show that the correct rate is obviously raised. High, it has a remarkable effect in the classification of traffic signs.

【作者单位】： 济南轨道交通集团有限公司;
【分类号】：U495;TP391.41
，

本文编号：1792082