Fingerprint Identification System Based On FPGA

Because of the characteristics of uniqueness，reliability and stability，the fingerprint hasbecome an important basis for personal identity authentication．Due to the advantages ofconvenience and efficiency，automated fingerprint identification system has been widely used inthe field of financial security，data encryption，e-commerce，and will play an increasinglyimportant role in our lives．The main work of this thesis is divided into two parts：fingerprint acquisition and fingerprintrecognition algorithm．As the deficiencies in real-time characteristics of traditional fingerprint acquisition system，an embedded fingerprint acquisition system design based on FPGA is proposed in thethesis．With the FPGA chip EP2C35F672C6as the processor and the MBF200as the fingerprintsensor to realize data transfer by the SPI interface，an embedded fingerprint acquisition systemwith high efficiency is constrcted，in which the off-chip SRAM is used to save the fingerprintimage data．Fingerprint recognition algorithm includes：pre-processing，feature extraction，featurematching．The feature matching algorithm is improved and a new segmented point patternmatching algorithm based on vector triangular principles is proposed in this thesis．Firstly，fingerprint classification is introduced as early matching to reduce the search area；Secondly，anew set of feature points is constructed with the center points，in which the second matching isdone by congruent triangles theorem and the reference points of the two fingerprint images (to-beidentified and template fingerprint image) are determined．Thirdly，the reference points as theorigin of polar coordinates，all feature points of new sets are mapped into polar coordinates bycoordinate transformation，and then the final matching is carried out by the variable boundarybox which can solve the non-linear deformation problem to a certain degree．Finally，statistics thematching pairs of points of the two fingerprint images and determine the matching resultsaccording to the setted threshold．The test shows that the hardware system design can run stablely and acquire fingerprintimage of good quality．Additionally，the fingerprint recognition algorithm used in the thesis canachieve each function to satisfy the design requirements．...