Research And Implementation Of Fingerprint Identify Algorithm Based On The ARM Embedded System

The fingerprint identification technology is a kind of important personal identification means, as well as the most widely used biological characteristic identification technology at present. Along with the developing of computers and microelectronics technology, fingerprint identification technology based on embedded is received more attention and investigation from people, causing embedded fingerprint identification system widespread applied in the fields of financial security, e-commerce, and safety inspection etc.This paper is discussed designing and implementing process of the fingerprint identification system based on the ARM embedded system, focused on the fingerprint identification algorithm which is adapted to the embedded system, detailed description of how to transplant algorithm in ARM embedded system, and analysed the experimental results.The algorithm adoptived in this task is mainly divided into three parts: preprocessing, characteristic extraction and matching. In preprocessing part, it is contained normalization, segmentation, binarization, denoising, thinning and so on. And characteristic matching algorithm is used a mathematical model which is irrespective with circumrotation and displacement. The subject is adopted EasyARM2200 embedded systematic developing platform, and usedμC/OS-II real-time operating system, thus completed the build task of a basic embedded fingerprint identification system.The system designed in this task has two kinds of working modes: the registration and the contrast. System is chose working way through the button after booting. While registration, the fingerprint sensor captured fingerprint image and preserved to the memory, the fingerprint recognition algorithm transplanted on EasyARM2200 to completed the digitized processing of the fingerprint picture, and kept formational characteristic samples in the end; similar to the registration, in contrast mode, the only different step is illegible comparing the new formational characteristic sample with the old sample which is already held, and then displaying the result on the digital tube monitor.The experiment on system developing platform indicates the fingerprint identification algorithm that should be adopted systematically has better practicability, transplantation and identification efficiency, and could distinguish and refuse the fingerprint images are not satisfactory collection. Through the parameters adjusting, system may achieve certain false acceptance rate and false rejection rate, and can satisfy the basic demand of the field of civilian fingerprint identification.