基于颜色模型和区域特性的交通标志提取方法研究

为了提升交通标志的检测效率,研究了基于RGB归一化交通标志阈值分割算法和基于HSI颜色模型的交通标志阈值分割算法,对比分析了两种分割算法的性能。针对分割后二值图像交通标志虚警率高的问题,研究了标志的区域特性,提出了基于区域特性的交通标志提取阈值处理方法,为进一步提升基于形状特征或基于机器学习的交通标志检测效率奠定了坚实基础。

Research of Traffic Signs Extraction Method Based on Color Model and Regional Characteristics

In order to improve the detection efficiency of traffic signs, we study the traffic sign segmentation algorithms based on RGB normalization thresholding and HSI color model thresholding, and analyze the performances of the two segmentation algorithms. In view of the high false alarm rate of traffic signs of binary segmentation image, we research the regional characteristics of traffic signs and present a signs extraction method based on regional characteristics. Experimental results verify the effectiveness of the method, which will lay a solid foundation for the further improvement of the efficiency of signs detection based on shape feature or machine learning algorithms.