Research On Automatic Fingerprint Identification System And Its Embedded Implementation

With the uniqueness and stability of the fingerprint, the fingerprint technology has become one of the important identification technologies. The automatic fingerprint identification system based on embedded system gradually becomes the focus of research for its miniaturization, intelligent and price advantage. So the main content of this paper is improving the automatic fingerprint identification system and realize it on embedded system on the basis of previous studies.In the image preprocessing aspects a segmentation method is implemented, then calculates the direction and frequency of the fingerprint image and uses these information to enhance the image. This paper also implements one binarization method of fingerprint images based on the 2nd derivative. Using the OPTA thinning algorithm, we can get a single pixel width binary image.This paper extracts features on the binary fingerprint image, and then removes the false feature structures with the orientation tracing method. After this post-process the features that has been extracted in order to eliminate the false features.Fingerprint image matching is an important technology of fingerprint identification system. This paper implements one fingerprint matching algorithms based on global star structure. Due to the use of the global features the rate of identification improved.Finally this paper transplants the automatic fingerprint identification system on the embedded Linux system whose hardware platform is ARM. After experiment, the algorithm that this paper has implemented can well realize their respective functions.