一种改进的SFIM高光谱图像融合算法

基于平滑滤波的亮度调节(SFIM)图像融合技术具有算法简单、计算快捷、光谱保真度高的优点,但同样存在边缘模糊、空间纹理信息提高不足的问题。针对上述问题,本文从瞬时视场角的角度出发提出一种改进SFIM算法,采用隔行隔列采样技术代替原SFIM算法中的逐行逐列卷积方法,从原高分辨率图像中计算模拟低分辨率图像。试验结果表明改进算法在边缘清晰度和空间纹理信息等方面均有明显的提高,且保持了原SFIM算法计算快捷的优点。

An Improved Smoothing Filter-based Intensity Modulation Algorithm for Hyperspectral Image Fusion

The smoothing filter-based intensity modulation（SFIM） technique is a simple,fast and high spectral preservation algorithm.But there are also some disadvantages of SFIM,such as edge blurring and the insufficient information improvement in textural information.To solve these problems,an improved SFIM algorithm is proposed from the point of view of the instantaneous field of view（IFOV）.This method can keep the advantages of SFIM such as the simple implementation and the less computational complexity.In order to derive the simulated lower resolution image from the higher resolution image,the proposed algorithm applies interlaced sampling to the convolution operation instead of progressive sampling.The experimental results prove that the improved SFIM method can effectively solve the problem of edge blurring and enhance the spatial textural information.