摘要
为了解决由于采集身份证图像时光照不均匀给二值化造成的困难,将基于概率的数字图像光照不均匀校正算法应用于对身份证图像的光照校正。该算法在对一个区域中的背景进行提取时,采取统计的方法,先计算该区域中像素亮度的均值和标准差,然后根据统计值进行背景灰度计算。这样有利于排除噪声点的干扰,鲁棒性更强。采用水平投影确定字符所在行的区域后,利用垂直投影确定单个字符的位置。实验验证了所提方法能有效地进行二值化并准确分割出身份证上的字符。
In order to solve the difficulties of binarization caused by uneven illumination of the acquired images, the digital image uneven illumination correction algorithm based on the probability is applied to illumination correction for the ID card. This algorithm uses a statistical method when extracting the background. It first calculates the mean and standard deviation of the pixels＇ brightness, then calculates the grayscale of the background, according to the statistical value. This is conducive to exclude the interference of noise points, and enhance robustness. The horizontal projection is used to determine the line position of the character, and the vertical projection is used to determine the position of the single character.Experimental results show that the method can effectively binarize and accurately segment characters on the ID card.
出处
《计算机工程与应用》
CSCD
北大核心
2015年第14期165-169,共5页
Computer Engineering and Applications
基金
国家自然科学基金(No.61300032)
关键词
身份证
图像分块
光照校正
字符分割
投影
ID card
image partitioning
illumination correction
character segmentation
projection
