Design And Implementation Of Fingerprint Identification Algorithms Based On ARM

With the development and breakthrough of information technology and integrated circuit technology,biometric identification technology has been developed rapidly in recent ten years.Biometric technology can be divided into face recognition technology,iris recognition technology,palmprint identification technology,fingerprint identification technology and finger vein recognition technology according to t he collected biological information.There are different performances in the aspects of collection difficulty,user acceptance,anti-counterfeiting performance and scheme realization.At present,fingerprint identification is the most widely used in the domestic and foreign markets,and has been highly recognized by a large number of users.It has the following advantages : small volume,low cost,low power consumption and ea sy development for embedded fingerprint identification system.A hardware platform of fingerprint identification system is designed by using the STM32F405RGT6 chip with ARM kernel and capacitive fingerprint sensor BF5333 A,firstly.And a lot of fingerprint data are collected using this system to study fingerprint recognition algorithms.Since the process of fingerprint acquisition by this system is highly consistent with the actual situation of users using fingerprint identification products,the fingerprint data collected in this paper can provide a good experimental basis for the subsequent algorithm design and verification.Then,the principle of fingerprint image processing is analysed deeply,based on the traditional fingerprint image processing method,and a large number of fingerprint images collected,including fingerprint image segmentation,enhancement,binarization and thinning.An improved fingerprint feature structure is proposed,based on the common fingerprint feature point structure.On a circle whose center is the location of feature point and the radius is a certain pixel length,three points are taken as the feature extension points uniformly.This structure is beneficial to increase the success rate of feature matching and reduce the error matching rate of feature points.The fingerprint identification algorithm is transplanted to the ARM hardware platform,with the achievement of fingerprint image acquisition and processing,fingerprint feature extraction and storage,fingerprint matching.And design the operating interface of fingerprint identification system,which includes the functions of registering fingerprint,verifying fingerprint and deleting fingerprint,by using Matlab GUI on the computer.The interface is connected with the ARM hardware platform through a serial port,realizing the interactive function of the fingerprint recognition system.The part of system testing includes the test methods and contents.Test results show that the design of this issue fingerprint identification system is stable,reliable and real-time application which reached a satisfactory outcome.At the end of the thesis,it proposed some questions which met in the work,and made a simple summary about some future research area.