Research Of Acquisition And Recognition System Based On Hand Features

As people’s emphasis on identity security growing, each security measure has beendeveloped. Biometric identification is welcome as a new method because of the advantagesof easy-carrying, confidentiality, stability and living-identification. Using multi-modalbiometric recognition can improve the recognition rate and increase users’ acceptancecombining with non-contact collecting method. It can be widely used in attendance andsecurity.This paper designs a non-contacted hand three-modal features acquisition system.Analyzing the imaging theory of palm vein, palm print and palm-shape, designs and selectsthe image sensor and lens which can obtain images under multispectral, also designs thelight system, hand limited area system and the PC interface. The system can obtain highquality images of palm vein, palm print and palm-shape and practically executerecognition with fusion algorithm and good man-machine interface. It solves the issue ofsensing diseases and low recognition rate of single-modal image deformation caused by thetraditional acquisition system.This paper also designs a system which can acquire and recognize hand featuresimage and give users results by LCD monitor and speaker with the core processor ofTMS320DM642. The hardware consists of infrared sensor module, light source module,acquisition module, display module, voice module and power and reset module, while thesoftware consists of the drive part to the related hardware, hand limited area module aswell as recognition algorithm module. Design the LED light source module which cancover the palm area evenly and completely with the AMC7150constant current driver chipcontrolling light switching and detecting palm position by infrared sensor module.Acquisition module uses the OV7620digital image sensor to acquire hand features imagesof640*480by reasonable configuration of registers. Display module uses the SAA7121Hvideo encoder chip and analog monitor to display the hand images and recognition resultscontinuously. Voice module uses WT588D-U chip working in one-line serial mode. Whensystem works, it will match the captured features images with the registration ones for comparison after preprocessing and give the recognition results back to the user. Thesystem builds the ideal hardware platform for non-contact hand multi-modal informationfusion recognition technology.