Based On Dm6446 Embedded Iris Recognition System Hardware Design And Implementation

With digitization and networking of personal identity, authentication technologywhich is masafe, reliable and convenient becomes indispensable. Compared with thetechnologies of traditional keys, password, RF Card and etc, Inherent human irisidentification can not be lost or forgotten, and is difficult to copy or forge. Comparedwith fingerprint recognition, face recognition and other biometric technology, Inherenthuman iris identification has a higher safety performance, and it is more convenient.This article designs and sets up an embedded iris recognition system based on TI’sDavinci video processing architecture. Based on the system, the iris image acquisition,image video display, TFT touch pad and the transplantation of iris recognitionalgorithms are completed. According to the characteristics of the algorithm, thealgorithm was optimized. The iris recognition system has the advantages of small size,fast processing speed, low power consumption, security, and system scalability. Themain work is as follows:1. The architecture of the embedded iris recognition system is design. The coreprocessing chip of the system, TMS320DM6446, which is an ARM+DSP dual-coreprocessing chip, and the associated peripheral circuits constitute the iris recognitionsystem together. The functions of system are discussed.2. The hardware platform for the embedded iris recognition system is designedand built. According to the specific requirements of the embedded iris system and theexpected effect to achieve, a complete DSP circuit diagram and the correspondingperipheral schematic are designed, and the schematic design work of PCB layout,alignment and electrical characteristics inspection is completed as well. The bare metaldebugging of the modules in the circuit board is done.3. The driver of each module in the Linux operating system is completed andtested. The test results show that，video captured module in the Linux system cantimely and effectively obtain the iris image data, and never dislocates or drops theframes of image data. The problems of the LCD image flashing has been solved in thedriver design of LCD touch screen. 4. The iris algorithm was optimized and transplanted into the embedded irisrecognition system. Algorithm optimization made the average processing time of oneframe of the embedded systems reduce to less than400ms form2~3s beforeoptimization.5. Based on the system level testing of the entire embedded iris system, themodule parameters and algorithm parameters are adjusted to To meet the practicalrequirements.