Research On Non-contact Fingerprint Recognition Based On Single Camera

In order to improve the quality of fingerprint image fundamentally,this paper proposes a method of fingerprint acquisition based on non-contact.This method has great research significance and many advantages for improving the quality of fingerprint image.However,one obstacle for the new 3d fingerprint recognition system to replace the traditional 2d fingerprint recognition system is its large size and high cost,mainly due to the use of high-cost lighting system or multiple cameras.This paper attempts to solve these key problems of current 3d fingerprint identification technology by developing a 3d fingerprint identification system based on single image machine.The principle of non-contact acquisition and imaging is different from that of contact,so the effect of the obtained fingerprint image is different from that of the contact image.So this article aiming at the existing problem of non-contact collection of images using the corresponding algorithms,mainly reflected in the pre-treatment,further improve on the fingerprint image segmentation,using the maximum entropy threshold segmentation algorithm for fingerprint image background and foreground,but based on camera,fingerprint images were collected some knuckle part also confused in the prospect area,so this article embodied in the segmentation algorithm to the knuckle part of segmentation,reduced the feature points extraction of subsequent work number of false feature points extraction in so as to improve the matching rate.First,in the part of the image enhancement using homomorphic filtering to enhance the algorithm improve the fingerprint image of ridge and valley line between the problem of low contrast,but after a filter to enhance fingerprint image remains a little burr,so this paper pointed to an enhanced fingerprint image using the improved second enhance Gabor filter,greatly improving the quality of fingerprint.In fingerprint matching,the method of constructing triangle vectors based on detail points is used.Finally,the results are verified by Matlab.Verification results show that the non-contact fingerprint system based on single camera proposed in this paper and the corresponding algorithm used in this paper can effectively complete the application of identity recognition.Due to the improvement of fingerprint image quality and the corresponding algorithm,the fingerprint identification system obtained has effectively improved the accuracy of fingerprint identification and played a promoting role in the development of non-contact fingerprint identification.It has a good application prospect and great research significance.