Embedded Fingerprint Identification System

Fingerprint identification refers to an identification technique of a person ,based on his/her inherent fingerprint physical characteristic by computer. Due to the uniqueness and invariability of fingerprint, the fingerprint identification is more attractive , compared with traditional methods such as password and ID certificate. However, Most of traditional high performance automation fingerprint identification system is based on PC online application. Now because of the wide use of the automation identification system ,it is required to be easier to carry and use. Therefore embedded fingerprint identification system becomes a hot topic. Besides the advantages above , embedded fingerprint identification is also superior to online fingerprint identification in extensibility and stability.This thesis pay much attention to the use of fingerprint identification, and what we do is summarized as three points.Firstly, we sum up what other people did about embedded fingerprint identification technology, and investigate its application ,market situation and its trend .Then we introduce the characteristic of embedded fingerprint identification system , discuss the problems and steps met in developing this kind of system. Also ,we do some further research on theory of fingerprint collector and its parameter .After comparing several kinds of fingerprint collector , finally we can tell the user which is the better due to his/her requirement.Secondly, in this thesis , an embedded fingerprint identification system based on DSP(Digital Signal Processing) is brought forward , which perfectly combine traditional PC-based fingerprint process algorithm and embedded system . Because of several disadvantages of embedded system, we draw much attention to algorithm and hardware design . Giving the rate of performance and cost, we choose DSP ,which is high performance but medium price . Then the implement of fingerprint collector , power manager and memory is introduce in detail. When it comes to algorithm , we do a lot of tests, simplify the algorithm and guarantee the system reliability .Moreover, a ridge distance algorithm which is based on FFT(Fast Fourier Transform) ,is quite suitable to embedded system .In the beginning ,the fingerprint image is transformed to spectrum image via FFT ,then it is divide the spectrum image into n directions ,and in each direction we estimate ridge distance .So the average of these distances is the finally distance. Tests with some typical images show...