The Research Of Automatic Generation And Recognition System For Security Seal

The seal is always deemed as a symbol of status and behavior, and it is used inall aspects of social life. Crimes which were caused by counterfeit seal occurredfrequently with the development of computer technology and improvement ofprocess with seal. The appearance of the seal imprint verification system realizedautomatic management and recognition of seal and effective prevent counterfeitseal caused crimes.This thesis based on the research of seal anti-counterfeiting technology andautomatic identification technology, designs a security seal automatic generationand recognition system scheme. The main results are as follows:(1)Analysis and comparison of generating anti-counterfeiting texture of sealand put forward a method of generating texture lines with the feature obvious sinecurve elements. As security curve, the single sine curve has high robustness butpoorer Information hiding features. According to the shortcomings of the sine curve,we introduce baseline parameters and constitute three sine curves with differentamplitude and initial phase into texture element. The texture can embed into thefive-pointed star of seal or margin of the seal, thereby security the seal.(2)Analysis the content of detecting seal image and the influence of the sealimprint verification. In the progress of image preprocessing, we use the arithmeticof extracting of red part with HSI model to extract the seal image. Then use themethods of binary grey and noise removing to improve the quality of image. Finally,design algorithm to rough image registration and get a complete pure seal image.(3) Analysis and comparison of texture feature extraction algorithm. Design analgorithm with a characteristic parameter by extracting the seal texture curve,which can fast identify detecting seal image according to the characteristics of seal.Window extracts the element of seal texture. Extract the characteristic parameter ofelement through the transform with voting mechanism and Hough. Finally realizematching with the detecting seal. This algorithm with Strong robustness pans outthe rapid and accurate identification of seal.This thesis design and implement the automatic generation and recognition system of security seal. According to the characteristics of generating seal, werealize the fast identification of seal by recognition of seal texture. We verify theeffectiveness of the guarantee system identification through the experimentalresults as well as high time performance, which meet the needs of the automaticrecognition of seal.