Research And Implementation Of Bank Draft Identification Technology Based On Image Matching

With the full development of our economy,bank notes are used more and more frequently.According to statistics,in 2005 alone,the notes being issued by China have about 1.8 billion,and the amount involved is about 350 trillion.In recent years,the economic development is to promote the development of the use of notes and the big boom.In February 2006,the people's Bank of China decided to establish the check image exchange system(CIS)based on the image technology and cover the whole China according to the situation of our country and learning from foreign experience.However,the frequently use of the notes,it makes the bill fraud criminal activities more and more frequently.With the development of imitation seal technology developed,the technology of digital image,the authenticity of identification of the bill is increasingly prominent problem.This has a huge threat to the property safety of the banks and the clients.At the same time reducing the bank staff to verify the workload and reducing the bill identification checks the false detection rate and false negative rate is very important.Therefore,based on the computer automatically to the authenticity of bank notes for the identification is a very important practical significance and value of application.This paper is based on the blind identification technology of digital image for bank draft identification.The main idea of this paper is using open check of the original bill image and honor of check bill image acquisition,using image processing algorithms to identify the authenticity of.Firstly,according to the image acquisition angle,displacement and so on,the image is based on the edge feature points.Secondly,the image segmentation is segmented,and the image is firstly matched.The last is the key areas for the main digital amount and seal part identification.In the amount of the digital part,the Hu invariant moments early identification,followed by the use of shape features of the amount of precisely identification,the stamp part firstly use the centroid of the seal for early identification,then respectively using SIFT algorithm and the NCC algorithm for image exact matching with quasi,combined with the correlation coefficient method for precisely identification of stamp.The final purpose of the identification is the authenticity of the bank draft.In this paper,on the HP PC,the simulation of bank draft identification technology is carried out under the Matlab software platform.Experimental results showing that the algorithm of this paper is of good accuracy,and can meet the actual needs of bank work.