印鉴自动识别系统中图像处理与 识别的方法 研究

采用基于边缘二值图的区域内灰度变换算法，提高印鉴图像局部灰度的对比度，从而防止了印鉴图像局部灰度过线使信息大量丢失的现象；提出了封闭凸多边形图像的提取方法－种子扩散浮置实体算法，解决了将待验印鉴图像从背景噪声中提取出来的问题；提出用重心点和中心点计算偏差角度，解决了待验印鉴与预留印鉴对齐的问题，并以内差和外差的概念，判断印鉴的真伪。

A STUDY ON THE METHODOLOGY OF HOW TO DEAL WITH AND INDENTIFY THE IMAGES IN THE AUTOMATIC SEAL-INDENTIFYING SYSTEM

A corresponding series of methods and algorithms were used to solve the problems in the automatic seal identifying system. In order to increase the contrast between partial gray levels of the seal image, the algorithm of the gray level transformation in a certain area was used on the basis of the fringe binary image. By doing so, it can avoid the mass loss of the information of the light gray level in the seal image. The method of presentation of the closed convex polygon image which was also known as the Seed Spread Floating Entity Algorthm was put forward, so that the problem of presenting the seal image was solved. Moreover, it solves the problem of how to make the seal image waiting for examination fit with the obligated one by using the warping angle calculated out of the center and the center of gravity. In addition, the concepts of inside and outside differences were adopted to judge whether the obligated seal is true or false. The above methods and algorithm have got quite good results in practice.