基于 OMAP3530的嵌入式多光谱掌纹识别系统

摘要

We implement an embedded multispectral palmprint recognition system using dual -core OMAP3530 platform.We uses Palm Cade palmprint recognition algorithm , enhance the system anti-spoof property by integrating the palmprint with palm vein , at the same time improve system accuracy from 91.33% to 95.33% for visible light.When implementing the system on OMAP3530, DSP processor with powerful operation capability is employed to concentrate on the complicated feature extraction operation , and the ARM processor is engaged in image capture , peripheral control , task scheduling and user interaction .By integrating the image capture module and the central processing board, the designed system has high recognition rate , strong anti-spoof capability and possesses the performance of real-time processing.%基于OMAP3530双核平台实现一个嵌入式多光谱掌纹识别系统。采用Palm Code掌纹识别算法，通过融合掌纹与手掌静脉提高系统防伪性，同时使算法的准确率从可见光的91．33％提高到95．33％。在OMAP3530实现时，让具有强大运算能力的DSP处理器专注于复杂的特征提取运算，ARM处理器则负责图像采集、外设控制、任务调度及用户交互。所设计的系统集成图像采集模块和中心处理主板，识别率高、防伪性强并且具备实时处理性能。