论文部分内容阅读
针对道路交通指示牌的识别,提出一种自动识别方法。首先对采集的图像进行预处理,包括把采集到的原始图像进行基于颜色和基于形状的两次检测,把RGB图像模型用HIS颜色空间表示和图像形态学的运算;其次对图像中不同的封闭子区域进行标记,并去除不满足面积阈值的子区域,得到数据后,采用多蓝色直方图和最小二乘法特征匹配进行特征识别,可以有效地识别出变形、缩放、部分遮挡及污脏的交通标志,从而得到交通指示牌的含义。
Aiming at the recognition of road traffic signs, an automatic identification method is proposed. Firstly, the collected images are preprocessed, which includes the color-based and shape-based detection of the collected original images, the RGB image model represented by HIS color space and the operation of image morphology. Secondly, Sub-regions, and remove sub-regions that do not satisfy the area threshold. After obtaining the data, feature recognition is performed by multi-blue histogram and least-squares matching, which can effectively identify deformation, scaling, partial occlusion and dirty Traffic signs, which get the meaning of traffic signs.