| Along with biological features recognition technology's development,the high precision biometric technology is widely used in the identity verification more and more. Such as: fingerprint recognitions, iris scanning, retina scanning , sound ripple , palm prints , facial features recognition and so on.The technologies including fingerprints,retina,iris biometric recognition have relatively high credibility and accuracy.Because fingerprint gathering is relatively convenient and the hardware is easy to be realized,the fingerprint recognition algorithm is more mature than the other biometric recognition technologies.The fingerprint recognition has higher usability and feasibility in terms of the overall performances.To remove the noises from the image and turn it into a clear point chart,the image must be pre-processed before the fingerprint is recognized.The purpose of the pre-processing is to extract the correct fingerprint characteristics and enhance the image identifiability.Fingerprint pre-processing technology is the foundation of feature extraction.At the same time it's a key for fast and efficient fingerprint identification.This paper particularly introduces the algorithm realization of fingerprint pre-processing,including algorithm simulation and DSP Programming.Some key technologies of fingerprint pre-processing algorithm are studied such as normalization , filtering noise , segmentation , binarization , thinning . An improved method of Gabor filter is proposed in this paper.Other parts of fingerprint pre-processing have been optimized at the same time.Adopt fourier analysis method as our basic tool in the research of fingerprint image processing.All the algorithms have been realized using MATLAB tool in this paper.The work theory and inner structure of TMS320C5509 series DSP chips and Fingerprint Sensor MBF200 also have been described simply.Based on grasping the principle of fingerprint pre-processing,the efficient fingerprint pre-processing algorithm designs are presented. |
PDF Full Text Request