基于非抽样Contourlet变换的红外图像和可见光图像融合算法

Fusion of Infrared and Visible Light Images Based on Nonsubsampled Contourlet Transform

分析了非抽样Contourlet变换(nonsubsampled contourlet transform,NSCT)的原理和优点,针对红外图像和可见光图像的特征,提出了一种新的基于NSCT图像融合算法,应用NSCT对图像进行多尺度、多方向分解,低频子带采取区域能量加权法融合,带通子带采取区域方差加权法融合,最后将融合的系数进行NSCT逆变换得到融合图像.实验表明,与其它融合算法比较,该算法融合图像效果较好.

The advantage and principle of nonsubsampled Contourlet Transform(NSCT)are studied.Focusing on the features of infrared and visible light image with the same scene,a novel fusion algorithm is proposed based on NSCT.Firstly,the sources images are decomposed at different scales and directions,the low frequency subband coefficients and the the bandpass subband coefficients are obtained.Secondly,different fusion rules are applied in different subband coefficients.The regional energy fusion rule is applied in lo...