| Iris recognition is a new biometric identification technology,which is widely used today.Because of its uniqueness,reliability,and against artifice merits,iris recognition technology has been making fast development.Furthermore,it will be proved to widely apply.Therefore,how to develop an iris recognition system is the hot field.This paper proposes a new project on iris recognition system with one embedded kernel,which is the Digital Signal Processors.It will be designed as one embedded platform for real-time image processing,which is composed of DSP,CPLD,network interface,and CMOS Image Sensor.according to the need of the iris image acquisition system,TMS320DM642 DSP of Texas Instruments is chose as the main image processor which is designed for video application specifically.We also chose OV7120 CMOS camera as the image sensor in the iris image acquisition system.The iris image acquisition system described in this paper is based on the TMS320DM642 core board.Through the extended interfaces of the core board a daughter card is designed by us,through which the communication between TMS320DM642 core board and OV7120 is accomplished.With the special functions of video port and EDMA in DM642,DM642 is connected with image capture system.The image can be captured without the support of CPU.All of driver and algorithms program were developed in Code Composer Studio(CCS)Integrated Development Environment offered by Texas Instruments(TI)and implemented in the system hardware.The software of iris image acquisition system includes diver module,image acquisition,processing,network module.Based on the module design strategy in software,we can replace the processing algorithm module easily and maintain other parts of the whole program unchanged.Furthermore,NIR ring LEDs source in iris image acquisition was designed.Modeling of iris image acquisition system for optimizing illumination schemes was built up and simulation experiment was made for better illumination uniformity. |
PDF Full Text Request