一种新的人民币号码图像分割方法

A Novel Segmentation Method of RMB Numbers Image

针对人民币号码图像前景与背景分离问题,在对现有的几种分割方法进行分析实验的基础上,提出了一种新的人民币号码分割方法,即基于形态学处理的直方图阈值分割人民币号码图像的方法,对目标和背景混杂在一起,阈值选取困难的图像,先进行形态学变换,处理后的直方图阈值就很好选取了。实验结果表明,该方法的错误分割率明显低于普通的直方图阈值分割方法。

With reference to segmenting RMB number image,some segmentation algorithms are proposed by scholars.After a few present segmentation algorithms are analyzed,A novel segmentation algorithm for RMB Numbers images is proposed in this paper,that is,the algorithm based on morphological histogram threshold value is proposed for segmenting an RMB number image.For foreground blocks merging into background blocks,RMB number images whose threshold value is difficultly selected.The morphological transformation is performed,and the processed the threshold values of histograms are selected easily.The experimental results have verified that the proposed segmentation algorithm is superior to the other histogram threshold segmentations.