The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B7. Beijing 2008 
The method need not re-sample the hyperspectral images, keep 
the spectral information unchanged, induced the 
undetermination of pixel decomposing, showed optimistically 
portend. The AMIRS system gives a opportunity to integration 
application of hyperspectral information and high special 
resolution images. 
The actualization of the workflow simple and crude, only 
simple algorithms are adopted, the fusion results can be 
zchieved but a lot has to be done to improve and test the 
method. 
5. REFERENCES 
Figure 7 result after pure pixel diffusion 
Ma Yanhua ,Wang Jianyu,etc, A New Vector Filter Based on 
Background Detection for Spectral Image, J . Infrared Millim. 
Waves,July 2006 
Weiming Xu, Jianyu Wang, Rong Shu,etc, A Airborne 
Multidimensional Integrated Remote Sensing 
System ,ISPRS,2008, to be published. 
Figure 8 result after one step pixel 
Decomposing and diffusing 
In figure 9 another fusion result is showed.The raw 
hyperspectral image is of 130*70 and the high spatial image is 
of 670*700. The left image is the raw hyperspectral image 
which is zoomed out; The middle one is the hyperspectral 
image after PCA fusion; And the right one is the image 
processed with the fusion method of this article. 
Figure 9 another experiments 
The results are spectral images ,with the spectral , it’s easy to 
produce classified maps. 4 
4. SUMMARY AND CONCLUSIONS 
In this article a type of hyperspectral image fusion method is 
bring forward, its workflow is introduced and realized » In the 
hyperspectral image fusion flow, the edge information and the 
radiation correlation between hyperspectral images and high 
spatial resolution images is utilized to induce the demand for 
registration precision that hard to achieve«, Experiments results 
with real remote sensing images are given« 
1178