In recent year, remote sensing technology has a rapid development. Moreover, the spatial resolution of satellites is greatly improved. Image fusion methods can combine low spatial resolution multispectral image and high spatial resolution panchromatic image to achieve high spatial resolution multispectral image. In this study, we fuse SPOT 10-m resolution panchromatic and 20-m resolution multispectral images into 10-m multispectral image. However, the images collected on different dates yield different fusion results for the same material because of the atmospheric condition. This poses a problem for change detection of the ground materials. Therefore, we propose a fusion technique with relative radiometric correction by color transform algorithm which matches the spectrum of the same material in one image to that of the others. Compare the fusion results, the changes of the land covers can be easily observed. In our experiments, we will show the application of detecting the changes caused by natural disasters.