Fusion and classification evaluation of SPOT-5 panchromatic and multi-spectral image

SPOT-5 multi-spectral image is rich in spectral information. Its four spectral bands are red, green, near-infrared and short-infrared, which have a good value in the area of forest vegetation application. SPOT-5 panchromatic image has high spatial resolution of 2.5 meter, in line with the requirements of the general application. So the application of SPOT-5 satellite data in optical image is very extensive. In this paper, by the software of ERDAS, SPOT-5 high-resolution and multi-spectral image are fused with method of PCA (Principle Component), Multiplicative and Brovey-transformation. The fused images retain multi-spectral information while maximizing the spatial resolution, which can not only improve the visual effects of image, and also can be used in classification. It provides professional graphics for the measurement and favorable data for remote sensing image interpretation. Classification is carried out to evaluate the effect of fusion with three different methods. And classification results of fusion images are compared with thee classification of original SPOT-5 multi-spectral image. The conclusions are: among the three integration methods, Brovey-transformation has the best visual effects, and its fused image retains better spectral information of the original image while maintaining sharpening the image effect. The classification accuracy of Brovey-transform fusion reaches 85%.