Embedded Fingerprint Identification System

Because of the fingerprint's stability and uniqueness, fingerprint recognition technology has been widely used. Now, with the performance improvement of embedded processor, automated fingerprint identification technology is in the direction of miniaturization. However, due to the high complexity of fingerprint recognition algorithm, the long response time of fingerprint identification system in an embedded environment makes people hard to bear. while, the use of high-performance embedded processors greatly increase the cost and power consumption, which prevents the application and popularization of automatic fingerprint recognition technology. This thesis improves some algorithms of fingerprint identification and studys some simplified algorithms, as a result, the automated fingerprint identification system can run smoothly in a mainstream embedded processor. And at the same time, study embedded platform based on ARM and LINUX, all these lay the foundations for the porting of fingerprint identification system. The main work of this thesis is as follows:1. In fingerprint segmentation section, after study the segmentation of fingerprint image based on gray and direction, this thesis proposes a new method which uses the mask of segmentation based on direction of point and traces the boundary of fingerprint image, and gets a satisfactory effect;2. In fingerprint enhancement section, after study the two-dimension traditional Gabor enhancement algorithm and its simplified decomposition algorithm, this thesis adopts the simplified decomposition algorithm and uses 12 directions to enhance an image, and achieves good enhancement effect;3. In fingerprint thinning section, this thesis compares several thinning algorithms, and uses look-up table thinning algorithm combining reservation templates of improved OPTA thinning algorithm, and gets good effect after noise weeding;4. Study the design and implementation of BootLoader, and achieve the purpose of burnning NandFlash by using an indirect method , and port Uboot to S3c2440;5. Study the design method of device driver under embedded Linux, and analyse usb device driver deeply, at last, program generic device functions and related structures of usb device driver according to experimental system .