摘要
设计并实现了一套基于TMS320DM6446和TVP5158的虹膜识别系统。硬件设计采用TMS320DM6446+TVP5158+NAND FLASH架构,以较低成本实现了两路虹膜图像的同时采集与处理。软件设计基于CodecEngine架构,ARM端负责视频采集、结果显示和数据库的管理,DSP端专注于虹膜识别核心算法,通过图像采集、虹膜定位、特征提取与特征匹配4个步骤来实现虹膜识别。该系统充分利用TMS320DM6446强大的图像处理能力,经过C语言级别和汇编语言级别的优化,该系统能获得较好的识别效果。
An iris recognition system based on TMS320DM6446 and TVP5158 is introduced. The hardware design is based on the TMS320DM6446 ＋ TVP5158 ＋ NAND FLASH architec- ture, which achieves the low-cost dual-iris image acquisition. The software design is based on the Codec Engine framework, where the ARM side is responsible for image acquisition, database management and user interface, and the DSP side focuses on the core algorithm, con- sisting of image evaluation, iris localization, feature extraction, and code matching. The sys- tem takes full advantage of the powerful image processing capability of TMS320DM6446. After code optimization under the C level and the assembly level, the system can obtain the high ac- curacy with the fast identification speed.
出处
《数据采集与处理》
CSCD
北大核心
2012年第6期658-664,共7页
Journal of Data Acquisition and Processing
