Research Of Fingerprint Matching Algorithm Based On Vector Triangle

Automatic fingerprint identification technology is a research hotspot in the field of biometric identification,according to the various technical problems,there are many domestic and foreign scholars have made in-depth research on fingerprint identification technology,and put forward a lot of algorithm or improved algorithm,and achieved fruitful results,but some difficulties still exist for fingerprint recognition in the fully resolved such as direction field accuracy is not high,the pretreatment effect is poor,slow matching speed and of low quality fingerprint image recognition effect is poor.The matching algorithm plays an important role in the fingerprint identification system.The matching algorithm is good or bad,is directly related to the accuracy of the matching results.At present,many researches on the matching algorithm,and has been put forward many successful algorithms,such as point pattern matching algorithm based on ridge structure matching algorithm.Based on these algorithms,the image quality must meet certain conditions,That is to say the image quality is not too bad,can provide enough information for feature matching.The matching algorithm based on point pattern as an example,this algorithm requires the number of feature points in fingerprint images must reach a certain number,otherwise unable to match.However,in practice,sometimes we have to deal with the poor quality of image to deal with these images.Under special circumstances,we put forward the new challenge to the matching algorithm.This kind of problem based on the key techniques of fingerprint recognition fingerprint matching is researched deeply,and put forward a new matching algorithm.At present,consumer products more and more small,people require the security of mobile devices continues to improve,the fingerprint recognition system is applied to these small consumer products such as mobile phone,laptop computer,etc..So to design a smaller,faster and has important practical significance for embedded fingerprint identification system.On the basis of previous work,this paper has completed the following work:(1)the analysis of the algorithm principle and the advantages and disadvantages of the mainstream fingerprint image matching,including the point pattern matching algorithm of fingerprint matching algorithm based on fingerprint ridge structure of fingerprint based on the fingerprint matching algorithm based on correlation.(2)the vector triangle based fingerprint matching algorithm is analyzed in detail,pointing out the shortcomings of existing algorithms,which is of poor quality fingerprint image,the effect is not ideal.We propose an improved matching algorithm of vector triangle based fingerprint ridge structure,adding fingerprint minutiae in the first match two,the relative angle,angle difference and other information to build a variety of threshold to improve the initial matching,so as to improve the accuracy,in two,by using the adaptive threshold,further improve the correct rate of matching.The results show that the algorithm for low quality fingerprint image has better matching effect.(3)using Verilog language to design the hardware circuit of fingerprint recognition algorithm based on FPGA.Fingerprint image detection,fingerprint image enhancement,fingerprint feature extraction and fingerprint minutia match module is designed as a dedicated hardware circuit,the hardware implementation of the fingerprint recognition algorithm,improved fingerprint recognition algorithm speed.Design applied to a variety of embedded system in the framework of the United Simulation Platform of MATLAB and Modelsim,can be matching fingerprint verification function.