The Design Of Embedded Multi-modal Biometrics's Identify System

As the information age comes, identification technology based on biometric is increasingly showing its value. Biometrics uses the main physiological characteristics or behavioral characteristics to do personal identification. Traditional biometric identification systems are mostly single-modal system, but single-modal biometric identification has some insurmountable problems. Meanwhile, because the market requests portable recognition system, the embedded multi-modal biometric Recognition System is began to research.In this paper, ARM-Linux embedded multi-modal biometric recognition system is designed. The system consists of three parts:ARM core plate takes control and computing tasks; the system expanded board of ARM expands the functions and provides interface; camera LCD and other peripherals complete image acquisition and display.This paper describes the hardware design, ARM system design process, and explains the design and debugging process of various functional modules. For software design, cross-compiler environment is set up, the U-Boot is modified and transplanted, the embedded Linux kernel and the root file system are compiled and transplanted. At last, hardware drivers are desinged based on the Linux system. Moreover, image acquisition and display application program and camera switch programm are designed.Finally, the embedded system is established to do experiments, which result shows that the system architecture design is feasible. This paper established a relatively common embedded multi-modal biometric platform, which has laid a good foundation for the research of biometric identification systems.