Research On The Key Issues Of DSP Embedded Fingerprint Identification System

Fingerprint recognition is a very obvious advantage in biometrics,face recognition,iris recognition with respect to the biological recognition technology has a wide application,low cost and development time,relatively mature technology etc.The core of automatic fingerprint identification system is fingerprint recognition technology.Although the fingerprint identification technology has been mature,there are still some minor problems,which have some optimization and improvement spaceThe research background is DSP embedded fingerprint identification system.First,the thesis introduces the general situation of biometrics and biometric system,expounds the concept and structure of fingerprint recognition technology and embedded system.Then,the thesis introduces the process of fingerprint identification system.And the image quality assessment problem,image processing problem,and fingerprint recognition problem are put forward.On the basis of collecting and sorting out the fingerprint identification data,the above problems are solved.In order to solve image quality assessment problem,a new method is proposed which evaluate the quality of feature points of fingerprint images.In the study of the irrational allocation of DSP module storage space,the thesis studied the fingerprint block processing,the fingerprint image filtering enhancement,the mixed programming interface rule and the call function transfer respectively.In the optimization of fingerprint identification algorithms,the region growing theory is used which can optimize the step of fingerprint image segmentation.the segmented image is processed by morphological method which includes corrosion and expansion,it can improve the quality and efficiency of fingerprint image recognition.The system is tested\The test results show that the DSP embedded fingerprint identification system proposed in this paper has a great improvement in performance and accuracy.