Use Of Optical In Vivo Fingerprint Scanner Fingerprint Image Preprocessing Algorithm

In field of fingerprint recognition, so many researches and experiments are made by scholars home and abroad that plenty of algorithms very useful in this field are proposed and a set of stable process are formed. But most of them are based on fingerprint images of NIST(National Institute of Standards and Technology)which are traditional ink printing. Because of background noise, a lots of pre-processing algorithm are generated.The primary topic of this thesis is the pre-processing algorithm of fingerprint image. In experiments, we found that it's very easy to get rid of background noise from images acquired with optical fingerprint recording device of our own. research on popular pre-processing algorithm based on direction image are made deeply.It is a key of result using pre-processing algorithm of direction image that how to get a more precise direction image. We propose a kind of algorithm combine Bresenham's draw line method in graphics to seek adjacent point position with using difference of gray seek direction image that improve the precise of seeking direct image.To refine useful information and erase background noise need divide a fingerprint image into segment of fingerprint and background. With quality of fingerprint image acquired by our own recording device, discussion on algorithm of segment based on variance of gray and direction image and how to practice a valid segment are made.Meantime,we found a way to judge quality of fingerprint image by using its segment result.Traditional image enhance and smooth algorithm can enhance useful information and smooth noise of image. But we erase most noise at time we acquired fingerprint image. Because of this, they are badly effective. For the reason of care feature of fingerprint algorithm using direction filter do an effective processing. We made a sample analysis of this algorithm.Binary is an important step of pre-processing. It is feature of finger ridge rather than gray of image that are focused on. Using black and white can not only show all feature of image but also be in favor of doing more processing by computer. For this reason, we discuss some algorithm valuable at point of mine and explain the algorithm we adopted in practice.Degree of reliable feature of fingerprint is depended on algorithm of thin. In the found of analysis of experience of forerunner, we brought forward a new algorithm. Using result of ridge center by scan fingerprint image from eachdirection, thin image with direction image. Further effective results are made by this algorithm than by tradition.There are many reason lead to false minutiae of thin image, which are spurs, holes, bridges, short lines etc. Such must be eliminated as many as possible. We found an algorithm using direction image for improving thin image.Through a serial of tests made in little quantity, we give the relative data as a valuation of our algorithm.