票据识别系统数据分割与提取

在手写体票据识别系统研究过程中，从带表格及噪声的背景中获取数据字符是至关重要的环节。实践证明［１］，阈值图象分割方法是一种简单而有效的数据获取方法，该文以国家统一发票为研究对象，在样本直方图分析的基础上，讨论了阈值选取和样本明暗度之间的关系．并由此设计了一种新的票据数据获取过程结构。与通常方法比较，将使系统整体速度有明显的提高。

The Data Segment and Extraction of CRS

In the study of checking pattern recongnition,to obtain the check data which is in the chart and noise background is the key procedure of the CRS (Check Recognition System).Usually the threshold method is a sim- ple and valid way to segment object image.Taking chinese common check digital images as our study object, this paper designs a new procedure about the check data obtaining which is based on the analysis of the his- togram of the check image.Testifted in practical CRS development ,the result of higher speed of CRS is obvious.