The Research And Implementation Of Fingerprint Image Segmentation Algorithm

Automatic fingerprint identification system(AFIS)has been widely used in all walks of life,people on its application in the industry is also easy to accept,study and use of fingerprint recognition technology has also been a lot of attention.With the continuous progress of society,there are more and more requirements of the work of identification,although,automatic recognition system has become a relatively mature method of identification,fingerprint identification system is used more and more widely,but the problems are increasing,fingerprint recognition fingerprint recognizer in recognition speed and recognition rate and stability is difficult to meet the actual use requirements.Fingerprint image preprocessing is the precondition of the whole fingerprint identification system,and the good results of pre-processing of the fingerprint image can reduce the operating time and improve the accuracy of the recognition.This paper for fingerprint image preprocessing of fingerprint image segmentation technology to do in-depth research,introduced and improved several classical fingerprint image segmentation algorithm,and simulation experiments are used to prove the correctness and effectiveness of the improved algorithm.The main research contents of this paper are as follows:Firstly,this paper,first of all,introduced the segmentation algorithm of fingerprint image based on Prewitt operator,the eight direction template and fingerprint image was convolved with the fingerprint image edge detection,which split from the foreground area of the fingerprint image.Second,introduced the segmentation algorithm of fingerprint image based on Canny operator,the algorithm of fingerprint image pixels using non maxima suppression,the fuzzy boundary became clear,was conducive to the next step of segmentation,and used a dual threshold segmentation,the segmentation effect better.Third,introduced an improved Gabor filter for fingerprint image segmentation algorithm,the segmentation algorithm is first of the fingerprint image using Gabor filter to filter.Then,in the segmentation of fingerprint images using the separation and merge algorithm,effectively to separate the foreground areaof the fingerprint image from fingerprint image.Secondly,the paper proposed a new algorithm of fingerprint image segmentation based on the gray level co-occurrence matrix with adaptive threshold.In many times after the test founded using the whole fingerprint image contrast square difference of mean(vM)as adaptive threshold segmentation,not only avoided prior algorithm manually select threshold tedious and imprecise problems,but also solve the this algorithms can't be applied in automatic fingerprint identification system problem.At last,in addition to the improved fingerprint image segmentation algorithm,the paper also proposed a fingerprint image segmentation algorithm based on the combination of Harris corner energy detection using adaptive threshold and mathematical morphology.The improved Harris corner detection calculating formula of threshold segmentation,by the total number of fingerprint images to detect the foreground region in Harris corner,and calculated the sum value of all Harris corner energy,to obtain the energy value of the average angle of the effective area of fingerprint image,as the threshold of the fingerprint image segmentation.After that,the basic operation of mathematical morphology was used to filter the image after segmentation.Finally,from the experimental results,the segmentation effect was obviously better than the existing algorithms,and it also proved that the improved Harris corner detection method is reasonable and effective.