基于小波包和Niblack法的枪号图像二值化算法  

Gun Code Binary Image Algorithm Based on Wavelet Packet and Niblack Method

作　　者：申森 李艾华[1] 姚良[1] 马治明[2] 

机构地区：[1]西安第二炮兵工程大学五系 [2]中国人民解放军96111部队

出　　处：《光子学报》2013年第3期354-358,共5页Acta Photonica Sinica

摘　　要：针对枪号图像字符易磨损、对比度太大或太小的特点,提出了小波包滤波器浮动阈值和Niblack法相结合的二值化算法.首先采用小波包滤波器对图像滤波,把得到的低通图像作为浮动阈值,并对图像进行二值化,得到第一幅待融合图像;然后采用Niblack法对图像进行二值化,得到第二幅待融合图像;最后把两幅图像按照灰度最小的方法进行融合,得到最终的二值化图像.实验结果表明该综合算法在二值化效果、适用范围等方面均有良好的表现,具有一定的推广价值.Aiming at the characteristics of the gun code image that characters are easy to wear and contrast gradient is too large or too small,a binarization method based on wavelet packet filter floating threshold and Niblack method is proposed.Firstly,the image is filtered using wavelet packet filter and binarized with the low pass image after filtering.Then,the image is binarized with the Niblack method to obtain the second fusing image.Finally,two fusing images are merged based on the minimum principle of image gray values to get the final binarization image.Experimental results show that the method has a good performance in binarization effect and adaptation range and can imply in many situation.

关 键 词：二值化 小波包滤波器 Niblack法 图像融合 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]