Research On Method Of The Security And Authentication For Bill Based On Multispectral Image

Payment of bills plays an increasingly important role in the national economy. At present, security and authentication of bills mainly rely on special paper of the bills, marks printed by special inks, stamp and watermark printed on the bills, and so on. When a bill is issued, image of the bill is stored in the database of the bank system, and while the bill is payed, the image of the bill is acquired, and compared with the image stored in the database of the bank system, the identification is implemented according to result of the comparison. This method requires the servers of the bank system be highly configured. With the extensive application of the payment of bills, the server of the bank system will be difficult to meet the demand of the identification of bills, meanwhile, the present method can not prevent cloning bills and altered bills effectively. This thesis researches a method of the security and authentication for bill based on multispectral image to solve the above problems.This thesis researches a method of the security and authentication for bill does not rely on the bank system, acquires a plurality of images in the same coordinate system of a bill by the bills image acquisition system, extracts the security features in each image, utilizes random information produced in manufacture of the paper for bills as element of encryption and security, forms bill fingerprint composed of colored filaments of the bill and information extracted from the bill content. When a bill is issued, encrypts the bill fingerprint via the RSA encryption algorithm, and generates a QR code utilizing the ciphertext, prints the two-dimensional code on the bill for anti-counterfeiting. When the bill is payed, decods the QR code, decrypts the decode content via the RSA decryption algorithm, thereby recovers the bill fingerprint. The identification is implemented according to result of the comparison of the fingerprint acquired form the image and the fingerprint acquired form the QR code.The thesis analyzes the method for acquiring multispectral image of bills and extracting security features from the multispectral image deeply. For features of the multispectral image, in a acquisition of multispectral image part, acquires image of color filaments of bills by improving light control and timing control of CIS, in a preprocessing of the image, proposes a algorithm for image edge detection based on a rectangular feature of bills, and standardizes the image of bills by correcting for inverted image and correcting for tilted image. In extraction of security feature section, discusses the method for localization security features of bills.Finally, the thesis designs the control processe for security and authentication of bills in detail, and designs the method for forming bill fingerprint by encoding colored filaments of the bill and information extracted from the bill content. By simulation for the bills image acquisition system, it shows that identification the bills via the method discussed in this thesis is highly reliable.