The encoding segmentation is the very important part of identification, the effectively segmentation will djrectly affect the recognition efficiency and speed. This paper proposed a segmentation based on feature of space and the vertical projection for characteristic of train's code. The experiments show that the segmentation is far superior the region growing and boundary tracking. The experimental results are demonstrated experimentally.%编码分割路货车编码识别过程中一个非常重要的环节,编码能否得到有效的分割将直接影响到识别系统的识别率和识别速度.针对铁路货车编码的特性,提出了一种基于铁路货车编码间隙特征和垂直投影特征的字符分割方法.实验表明,货车编码字符的分割效果远远优于基于边界跟踪法和区域增长法的常见字符分割方法,特别是在处理逆光、雨天等复杂情况下的图像时,此方法使识别系统具有较高的识别率和鲁棒性.
展开▼