Study on Data Fusion Algorithms of Landsat7 ETM+ PAN and Multi-spectral Imagery

Fusion of images with different spatial resolution can improve visualization of the images involved. This study tries to show that the fusion of the images from the same sensor system can improve quality of the original images. Four image fusion algorithms were used in the study of data fusion of Landsat 7 ETM+ imagery, taking southeastern part of Beijing City as the case study, they are the Smoothing Filter-Based Intensity Modulation (SFIM), High-Pass Filter (HPF) Transform, Brovey Transform, Multiplication (MLT) Transform. The effectiveness of the four fusion algorithms has been evaluated based on mean, deviation, information entropy, average gradient and correlation. The study reveals that the SFIM transform is the best method in retaining spectral information of original image, which does not cause spectral distortion and it has highest spatial frequency information. Therefore, fused images from the same sensor system can be used for improving visual interpretation and data quality.