The Research Of Ear Recognition Technology Based On NiosII

Biometrics recognition is making use of special physiology or behavior characteristic to realize personal identification. It provides way of biometrics recognition with high credibility and high stability. Ear recognition is a new branch of biometrics recognition, also is one of the challenge topic in computer vision and pattern recognition. The special physiology position and structure characteristic of ear makes ear recognition can be extensively applied at public safety and information safety filed. And more and more people start paying attention to ear recognition. The research of ear recognition currently still be placed in an experiment test stage, there are many problems to be solved for developing a practical ear recognition system.The research points of this paper particularly mainly include: ear matching algorithms, the design of hardware platform and NiosII, uClinux operation system transplantation.For the feature extraction of human ear, this paper proposed a new approach by using improved 2DPCA to reduce the feature dimensions, which lead to poor real-time capability and lack of data storage space. First of all, pre-processing of human ear pictures has been completed. Then an improved 2DPCA algorithm was used to compress feature dimensions. Finally, BP neural network is used to classify ear. Experimental results show that this method has the advantage of real-time and reduction of feature data, and also maintains the recognition rate.As aspect in the hardware platform, the function request of ear recognition hardware system has been researched, and a built-in hardware platform of NiosII is designed, which includes external memory and serial-communications interface.The uClinux operation system and the structure characteristics of document system have been studied. The measures of kernel reduction and compiling and the workflow of Bootloader program are discussed.The effect of ear characteristic matching algorithms is good on MATLAB. The building of hardware platform and transplantation of uClinux operation system have been finished. Then an ear recognition program is running on uClinux. The research results lay the foundation of embedded ear recognition.