多灰度图象中信息提取及二值化

Information Extraction and Binarization in Grayscale Images

研究了黑斑、阴影、污迹等污染的多灰度文本图象，对其进行识别提取，得到清晰的二值文本图象，图象的二值化的实现有利于图象的分析、ＯＣＲ的处理、支票图象处理和图象传输．研究了ＹＤＨ算法，考虑了图象的局部信息，采用动态局部阈值法，并在ＰＣ－４８６计算机上实现了ＹＤＨ算法思想．

The extraction of binary images from grayscale document images with background pictures, shadow, highlight, smear and smudge is a common critical image processing operation,particularly for the document image analysis,optical character recognition (OCR),check image processing and image transmission and understanding. The paper provides a technique forbinarisation and systematically evaluates and analyses the algorithm by Yasuda,Dubois andHuang,in which the local area character information is considered and the dynamic local areathreshold method is adopted. This algorithm is successfully implemented on a PC 486 andsome good binarization image results are obtained.