基于NSCT变换的红外图像增强新算法

针对夜间红外图像噪声大、对比度低的问题,探讨一种基于正态分布特性和NSCT变换相结合的夜间红外图像增强算法。充分利用了高频子带的系数标准差、系数均值和每个子带系数最大值的构造自适应阈值,保护图像的边缘细节,抑制图像噪声;提出一种新的弱边缘系数增益函数,有效地提升了图像的弱边缘信息。低频系数采用了基于正态分布特性的对比度拉伸增强算法。

Infrared Image Enhancement Algorithm Based on the NSCT Transform

Because of the low contrast and loud noise of night infrared image,the infrared image enhancement meth-od is presented in this paper,which is based on normal distribution characteristics and NSCT transform. By making full use of the high frequency sub-band coefficient of standard deviation,coefficient of mean and each sub-band coefficient maximum value for the construction of adaptive threshold,the image edge details are protected to better suppress the image noise. It also puts forward a new weak edge gain coefficient function,so weak edge information of the image can be effectively enhanced. The low frequency coefficient is based on the normal distribution charac-teristics and the contrast is enhanced.