【作者】 闵晶妍；

【导师】 苑玮琦；

【作者基本信息】 沈阳工业大学 ， 检测技术与自动化装置， 2005， 硕士

【摘要】 指纹的不变性和唯一性使指纹识别技术成为目前应用最广泛的身份验证与识别技术之一。在犯罪嫌疑人的识别(AFIS)系统、门禁系统、计算机安全、网络安全、考勤等方面有着显著的应用。指纹图像预处理是指纹自动识别过程的第一个处理环节,它的好坏直接影响着自动识别系统的效果。 本文通过大量的实验,综合运用图像处理和模式识别技术,深入、系统地研究了自动识别的理论和技术,建立了自动指纹识别预处理系统。针对目前研究现状,对一些关键技术问题进行了深入探讨,提出了一些新的方法。全文内容包括指纹图像分割、纹线提取、细化等几个步骤。本研究做的具体工作如下: (1)提出了一种基于灰度值频数和十二方向的指纹图像分割算法。该方法的阈值可以根据图像自然决定,避免了人为选择阈值的困难和不准确性。对脊谷线灰度值相差较大、不大、较小的指纹图均能容易而准确地分割。实验表明,与常见的分割方法一方向图、方差法相比,该方法的平均误分概率大大减小,比方差法降低5.7875%,比方向图法降低5.6625%,且对指纹图像脊谷线的对比度和方向性要求不高,鲁棒性更强。 (2)有效地实现了指纹纹线的提取。本文对目前指纹纹线提取的算法进行了编程验证,并做了一些修改。提出了利用周围块的方向信息校正当前块的方向;提出并构建了八方向平滑和增强滤波器。实验结果表明,方向校正后的滤波图像更清晰,而且少了毛刺、断线、短线等伪特征点,提取的纹线更接近于实际的指纹纹线。 (3)实验并验证了一套可行的查表细化方法。实验结果表明,细化后的指纹图像能保持原始指纹图像的特征点,并去除多余的像素信息,保证只有单象素宽,节省了后续算法的处理时间。更多还原

【Abstract】 Fingerprint recognition techniques become one of the most extensive identification and recognition techniques currently because its invariability. It is used broad in AFTS system, the control of gate system, computer safety system, network safety system and check on work attendance system etc. The fingerprint image preprocessing is the first step of automatic fingerprint identification system. Its quality affects directly the result of automatic identification system.Based on a lot of experiments, using technology of image processing and pattern recognition, the theory and methods of automatic identification are investigated and the automatic fingerprint preprocessing identification system is developed. At the same time, the key problems are discussed and some new methods are presented aiming at the status now in this field. The main contents are fingerprint image segmentation, minutiae extraction, thinning etc. The main jobs are as follows:(1)A fingerprint image segmentation algorithm based on the frequency of gray levels and 12-directional image is presented. The threshold can be selected naturally according to the image in this method and the difficulties in selecting the threshold are avoided. The fingerprint images with any difference in ridge and vale gray levels can all be segmented accurately. Experiments show that the method is 5.7875% lower in p error than squares difference and5.6625% lower in p error than directional image method. Moreover, the method is more robust because the difference of ridge and vale gray levels, orientation information isn’t rigorous.(2)The method of ridges extraction is studied by experiments and amelioration. The direction of current block is changed according to the direction around the block, because the variety of fingerprint image direction is smooth. The eight-direction smoothing and enhancing filter are also proposed. As a result, the ridges can be extracted correctly.(3)The method of fingerprint image thinning is researched. Erasing the current point is decided according to the eight points around the current point. This method is simple and the results of thinning are very ideal.更多还原