基于NSCT和IFCNN的红外与可见光图像融合算法

Infrared and Visible Image Fusion Algorithm Based on NSCT and IFCNN

针对红外与可见光图像具有不同特征信息以及在融合中存在清晰度低、细节信息模糊的问题,提出一种基于NSCT和IFCNN的红外和可见光图像融合的新方法.首先对两幅源图像进行预增强处理,以提高图像的对比度;然后通过NSCT分解得到高频和低频子带,将分解后的低频子带进一步进行DWT分解,得到高、低频子带,其中在低频部分采用局域能量加权方法进行融合,在高频部分采用IFCNN卷积神经网络融合框架进行融合,采用小波逆变换对低频和高频进行反变换,获得NSCT重构所需的低频部分,再对NSCT分解后的高频部分采用IFCNN的方法对图像进行处理;最后对处理后的低频和高频子带通过NSCT逆变换获得最终的融合图像.试验结果表明,该方法在视觉和客观指标上都具有较大的提升.

This paper aims to address problems of infrared and visible images with different feature information,low definition and fuzzy details in fusion.A new fusion method of infrared and visible images based on NSCT and IFCNN was thus proposed.Firstly,two source images were pre-enhanced to improve the image contrast.Then high frequency and low frequency sub-bands were obtained by NSCT decomposition.The decomposed low frequency sub-bands were further decomposed by DWT to obtain high and low frequency sub-bands.local energy weighting method and IFCNN convolutional neural network fusion framework were used in the low frequency part and high frequency part,respectively.Low frequency and high frequency were inverted to obtain the low frequency part required for NSCT reconstruction by using inverse wavelet transform.Then,IFCNN was adopted to process the high frequency part after NSCT decomposition.Finally,the final fusion image was obtained by inverse NSCT transformation of the processed low frequency and high frequency subbands.The experimental results show that this method can be used to improve visual and objective indexes.