| With the rise of e-commerce, online banking and other new financial services, the demand for personal information’s security and automated authentication is increasingly prominent. Automated fingerprint identification is the current and mature technology of authentication with a wide range of applications. However, the traditional fingerprint recognition’s technology must use fingerprint scanner to get fingerprint image,and then upload to PC for processing.Its hardware is large size, does not have the portability. The combination of fingerprint recognition and embedded systems changes suitation.Building embedded fingerprint identification system with functional independence and compact size is precisely required in the current situation.This paper studies and designs the embedded fingerprint identification system based on arm and linux platform. First introduces the background of embedded fingerprint identification’s research,and then from the bottom to the up, discusses selection and design of embedded fingerprint identification systems’hardware, customization and programming of Bootloader, cutting and transplantation of embedded linux operating system, research and implementation of fingerprint recognition’s algorithm.In fingerprint recognition algorithm,this paper discusses the fingerprint image preprocessing, including segmentation, balance, convergence, smoothing, enhancement, binarization, thinning; also discusses the image registration algorithm based on feature points of fingerprint and the feature point matching algorithm based on the boundary box. During the studies the paper use mini2440development board from Guangzhou Friendly Arm company and the TFS-M74A semiconductor module from Ten Fingers technology company in Shenzhen. Experimental results show that the effect of embedded fingerprint identification system designed in this paper to recognize is good,and it meet the requirements that we want it to meet at the beginning of design. |
PDF Full Text Request