摘要
提出一种最大极值区域(MSERs)结合分水岭变换的粘连交通标志分离算法。首先通过MSERs对图像进行二值化,选取在多个二值图像中形状均保持不变的区域作为候选标志区域;然后,判断是否为粘连标志,若是则用分水岭变换分离;最后提取单个标志。实验结果表明,该算法能够有效分离粘连交通标志,提取到完整的单个交通标志。
A new algorithm for the separation of adhesion traffic signs based on MSERs and watershed transform is proposed.Firstly,the image is binaried by several threshold,the region that keeps unchanged shape is selected as the interest.If the region is a adhesion,watershed transform is used to separate the adhesion.Finally,the single traffic sign is extracted.The experimental results show that our algorithm can effectively separate the adhesion signs,extract single traffic sign.
出处
《微型机与应用》
2016年第13期48-50,54,共4页
Microcomputer & Its Applications
