摘要
针对指纹图像可能出现模糊、指纹区域过小、奇异点位置过偏等问题,提出了一种基于信息可用性评价与频谱分析的指纹图像质量增强算法。对指纹图像进行信息可用性评价,对不合格的指纹图像提示进行重新采集;对合格图像进行傅里叶变换并求取其频率均值和方差,计算指纹频谱图上内环、外环、中环的频谱能量值与去除直流分量后的频谱总能量值之比,以此确定指纹的清晰程度。对需要进行质量增强的指纹,利用圆滤波器去除其高频与低频干扰,利用方向滤波器连接断纹并去除粘连。实验结果表明,该方法能准确判断指纹图像的可用性,有效地增强指纹图像质量,并因其只对低质量指纹进行增强,故能有效提高指纹自动识别速度及准确性和可靠性。
It proposes a fingerprint image quality enhancement algorithm based on information availability and spectrum analysis in order to solve the problems which may appear in a fingerprint image, such as the image blurs, the effective area of the image is not enough and the positions of singular points are overdue. It evaluates the information availability of a fingerprint image, and re-captures the failed fingerprint images. It takes Fourier transformation on the qualified image and obtains its average and root-mean-square uncertainty of the frequency, then calculates the ratios of the spectrum energy values of inner ring, outer ring and the ring to the spectrum energy value of total energy after removing the DC components to determine the clarity of fingerprint image. For the fingerprint image which needs to be enhanced, round filter is used to remove high frequency and low frequency interference, directional filter is used to connect broken lines and remove adhesions. Experimental results show that this method can determine the availability of fingerprint im- ages accurately, enhance the fingerprint image quality effectively, and only low-quality fingerprint images are enhanced, so it can im- prove the speed and accuracy and reliability of the fingerprint recognition effectively.
出处
《计算机工程与应用》
CSCD
2012年第8期207-210,共4页
Computer Engineering and Applications
基金
湖南省自然科学基金(No.10JJ2048)
湖南省重点学科建设经费资助项目
关键词
信息可用性
指纹识别
图像质量评估
频谱分析
指纹增强
information feasibility
fingerprint identification
image quality evaluation
spectrum analysis
fingerprint enhancement
