Recently, the requirement for authentication has been get a much higher level as thedevelopment of information technology. And fingerprint recognition is one of the effective andreliable biometric methods. As it’s unique, stable and easy to collect, the application offingerprint recognition is much broader than other biometric methods. In addition, it iscomprehensive as the technology has reached accurate and quick level.In this paper, a solution about the automatic fingerprint identification system based onARM9has been given. And the system has been transplant into the Linux system. The softwareenvironment has been researched and the block diagram has been ploted, either. Besides, thedetailed algorithm flowchart of the fingerprint recognition has been given.In the process of fingerprint image, first of all, the M-FastICA algorithm has been used toextract the fingerprint image and remove the noise. It could improve the robustness of theidentification system. Secondly, the image has been preprocessed which is including gray-scalefiltering, image enhancement, image binarization and refinement. Then the fingerprint imagefeature has been extracted. At last, the fingerprint image has been matched base on a new PFSValgorithm.After tested, the algorithm could achieve a good matching level. The automatic fingerprintidentification system has a high rate of fingerprint recognition. Besides, the rejection rate anderror rate are also improved. It’s a safe, reliable and simple to expand system. |