The Research Of Fingerprint Recognition Arithmetic And Hardware Realization

As an important part of means of ensuring the security of military information, our armyhas put forward new and higher requirements for user authentication of information technologymilitary equipment. Only fast, reliable and correct certification function can lay a solid securityfoundation for information technology military equipment to carry out combat tasks. Thetraditional "user name+password" to achieve technical equipment certification can not reallyensure that the equipment user is a legitimate user. Therefore, the subject considers using theinvariance, stability and irreproducibility of the fingerprint to design and implement a new wayof information technology military equipment certification.Taking into account the harsh battlefield fingerprint acquisition environment and thereal-time requirements, the traditional AFIS (Automatic Fingerprint Identification System) cannot be directly used to implement the information technology military equipmentcertification.This thesis focuses on the problem of low-quality fingerprint image matching andhow to identify the fingerprint quickly. The specific research is completed as follows:The thesis has proposed a low-quality fingerprint matching algorithm based on multiplefeature vectors. The algorithm combines the fingerprint feature points with the textureinformation to build a new multi-feature vector; taking into account the characteristics of thelow-quality fingerprint, in the analysis of the characteristics of the RST (Rotation ScalingTranslation) model and affine transformation model, proposes to use the least squares method toestablish the most general affine transformation model instead of the RST model for low qualityfingerprint matching; taking into account the complexity of the application environment,proposes the beginning of the match and the final match of the two links to improve low qualityfingerprint matching efficiency.The thesis has proposed a fast fingerprint image recognition algorithm which can be easilyhardware implemented based on adjacent feature points. The arithmetic chooses the most criticalbinarization and refining in the pre-processing stage, and sums up the extraction method whichcan be easily hardware realized. In the final stage, in the base of analyzing the characteristics ofthe fingerprint topology, the arithmetic uses the adjacent feature point information to completethe last match of the fingerprint image.The thesis has designed a FPGA-based fingerprint image processing module forauthentication. The thesis has detailed the principles of the module composition and the mastercontrol program design, and through the device selection of fingerprint acquisition circuit unitand the central control unit in detail, proposed the communication process based on the SPIinterface.The fingerprint image processing module has implemented the fast fingerprint image recognition algorithm based on adjacent characteristics. For the image preprocessing, featurepoint extraction and fingerprint matching which are the key parts in the fingerprint imagerecognition algorithm, the thesis has proposed the programming methods in hardware. Finally, itpresents the functional verification and performance evaluation. Judging from the results, therapid fingerprint image recognition algorithm has achieved the ideal balance between therecognition performance and resource consumption.