Design Of A Biometrics Recognition System Based On Finger Sensing Capacitive Reactance Spectrum Measurement

The current biometrics recognition(BR)technology(such as fingerprint recognition,iris recognition,face recognition,etc.)is mainly based on image acquisition and processing method,which has various deficiencies.The early stage of the team for the first time put forward a new kind of bioelectrical impedance spectrum(BIS)measurement based on the palms of BR method,designed a type of frequency sweep BIS measurement system based on ARM.But in the experiment,due to external environment such as finger wet degree,contact pressure,BIS data repeatability is poor,and the recognition rate is low.On the basis of the previous studies,this paper designed an improved measuring BR system based on finger sensing capacitive reactance spectrum(CRS),in order to improve the measurement repeatability and recognition rate.In this paper,the main content is as follows:1.This paper introduces the principle of finger sensing CRS measurement,describes the finger sensing model structures,and selects the measurement scheme,and introduces the Multisine excitation signal synthesis principle and synchronization of the whole cycle sampling principle.2.A fingers sensing CRS measurement system based on combination of FPGA+STM32 is designed,in which the FPGA controls 14-bit high precision DAC AD9767 to produce Multisine excitation signal,and 14-bit two-way high-precision ADC AD9648 to fulfill dual channel synchronous integral period sampling on excitation signal and response signal,then data cached in the FIFO of FPGA are transmitted to STM32F429 to save sampling data,and FFT is processed in the upper machine,finally fast measurement of 31 points CRS in the range of 32-992kHz frequency is achieved.3.A tester’s finger sensing CRS data are collected and repeated experiments using the finger sensing CRS measurement system are made.The experimental results show that repetitive effect is very good,and external factors(such as finger humidity,temperature)have little impact on experimental repeatability;100 students and teachers in Xi’an University of Technology finger sensing CRS data acquisition experiment was carried out,and 100 sample library and test library are established.4.By using a method of Euclidean distance and support vector machine(SVM)method to conduct recognition experiment,recognition rate of Euclidean distance can reach 66%,and the SVM algorithm can achieve 84%,which verify feasibility and validity of the biometrics recognition of finger sensing CRS method.This paper lay the foundation for the development of instrument and large-scale experiment in the future.