纸币清分图像识别算法研究

针对国产纸币清分机在图像识别技术上,存在挑残能力弱和分类一致性差的缺点,提出了3种基于图像识别技术的算法。通过对采集的图像数据,采用五线找点算法对纸币大小进行判定纸币的面额;根据加权灰度值提取纸币图像的特征区域判正反;采用网格划分法统计背景点与纸币图像点的数量关系判定纸币的残缺率。

Research on the methods of Paper Currency and Image Identification

Concerning the shortages on Image Processing of the previous designs of Paper Currency Identification Instrument,such as poor at deformity seeking-out and classifying the rank and directions of new and used bills,a series of identifying methods are presented.By the collected database of images,use the Five-line Seeking-points method to analysis the size and determine the denomination;based on the characters of gray scale of images,take and analyze for judging the face and obverse;Moreover,the gridding method is taken to measure the relation between points on background and images for deformity seeking and the taint degree.