一种基于光照模型的多源遥感图像融合算法

A Better Multi-Sensor Image Fusion Method in Remote Sensing

多光谱图像和全色图像是目前卫星遥感领域最常见的传感器图像。为了充分地发挥这两类遥感图像数据的互补性信息 ,增强二者的清晰度和解译能力 ,在 SFIM算法的基础上 ,将 IHS变换与 SFIM相结合 ,并对原有的均值滤波器进行改进 ,提出了一种自适应加权均值滤波器。通过一组多光谱图像和全色图像的融合实验 ,并对比常用的 IHS融合方法和 SFIM方法 ,证明了新算法在保持多光谱图像光谱特性的同时 。

We analyzed SFIM(Smoothing Filter-Based Intensity Modulation) fusion technique proposed by Liu in 2000 and came to forming the opinion that SFIM can be improved in two respects: (1)combining SFIM with IHS(Intensity, Hue, Saturation) fusion technique;(2)smoothing filter different from that used by Liu . In this paper we explain in much detail how to combine the two multi-sensor image fusion techniques and also discuss in some detail the design of the smoothing filter--called by us the adaptively weighted mean filter--we propose. We performed simulation tests that fused multispectral images and panchromatic images with IHS, SFIM and our improved SFIM fusion techniques respectively. Simulation results show preliminarily that our improved SFIM method is better than IHS or SFIM method and can get excellent spatial resolution while preserving the spectral information of multispectral images.