摘要
针对纸币号码识别过程中图像分割存在的问题,提出了一种逻辑脊谷判断的图像分割方法。该方法可以直接从原始图像中提取号码线条,经过简单的迭代法二值化和数学形态学闭运算处理,分割线条清晰准确,算法简单,省去预处理步骤。用现有的图像分割算法与本文算法比较,本文算法处理速度快,分割图像准确。
In order to recognize numbers on paper currency, an image segmentation algorithm is presented based on logic spine judgment. Number lines in original picture can be draw out directly through the method, after the method of simple iteration binarization and morphological close operation, the lines image is clear and accurate. The method is simple and the preprocessing is omitted. Compared with the existing iraage segmentation algorithms, the experiment results show the new algorithm is rapid and veracious.
出处
《计算机测量与控制》
CSCD
2005年第8期827-829,共3页
Computer Measurement &Control
基金
辽宁省教育厅高校科研项目(2004C069)。
关键词
图像分割
纸币号码
逻辑脊谷判断
image segmentation
paper currency numbers
logic spine Judgment
