209 
most in comparison with other fusion method and its entropy is 
less than fusion images based on PCA and wavelet. The 
standard deviation of fusion image based on PCA is closer the 
standard deviation of original multi-spectral image TM ,which 
indicates this method has less distortion of spectral 
characteristics . 
In the view of entropy, the entropy of wavelet-based fusion 
image is greater than these of other methods. So the information 
contained in fusion image based on wavelet is greater, which is 
to say, the quality is better than these of other three methods. 
The fusion image based on Brovey is the least, and its mean is 
small, so the image based on Brovey has less information and 
the whole image is darker than images of other methods. 
The correlation coefficient reflects the redundancy degree of 
images. According the table 1,we know the original image 
contains a plenty of redundant information and the correlation 
coefficient of fused image based on wavelet integrated PC is the 
least, which indicates that this method has less redundant 
information than other method. 
Figure3 original image: TM (right) and SAR (left) 
Figure 4 the fused images based on (a) Brovey transform, (b) 
PCA, (c) IHS and (d) wavelet integrated PC 
According to previous statistics and analysis, we can draw a 
conclusion that the fusion method based on wavelet integrated 
PC is efficient for merging low resolution multi-spectral TM 
image and high resolution SAR image. Though other fusion 
methods have advantage in some ways, the fused image based 
on wavelet integrated PC is optimal when considering all 
aspects. This method integrates the advantages of wavelet and 
PC and the method based on wavelet has been applied to fuse 
SPOT panchromatic and multi-spectral images successfully. 
4. CONCLUSIONS 
In this paper, a new fusion method based on discrete 2-band 
wavelet was presented and several fusion methods are used to 
merge the multi-spectral TM and high resolution SAR image. 
Several parameter including mean, standard deviation, 
correlation coefficient and entropy were adopted to appraise the 
fusion products. At last the result of experiment proves that the 
fusion method based on wavelet integrated is more efficient for 
fusing the TM and SAR data in comparison with other fusion 
methods. First the redundancy is little and entropy is large. 
Secondly the fusion image preserves better the original image 
radiometry. 
Though the fusion result is better, this method still has some 
question due to the limitation of time and data, this paper did 
not experimented, only to discuss from the theory. In general 
biorthogonal (and symmetrical) wavelets are more appropriate 
than orthogonal wavelets for image processing applications 
(Strang et al, 1997). Biorthogonal wavelets are ideal for image 
processing applications because of their symmetry and perfect 
reconstruction properties. So the fusion based on biorthogonal 
wavelet is better theoretically. The further study is need later. 
REFERENCES 
Aiazzi, B., L. Alparone, S. Baroni, A. Garzelli, 2002. Context- 
Driven Fusion of Spatial and Spectral Resolution Images Based 
on Oversampled Multiresolution Analysis, IEEE Transaction 
on Geoscience and Remote Sensing, 40(10), pp2300-2312 
Chavez P. S. and Kwarteng A. Y., 1989.Extracting spectral 
contrast in Landsat Thematic Mapper image data using 
selective principle component analysis, Photogramm. Eng. 
Remote Sens., 55(3), pp.339-348 
Gillespie .A. R., Kahle .A. B and Walker .R. E., 1987 .Color 
enhancement of highly correlated images—II. Channel ratio 
and ‘chromaticity’ transformation techniques. Remote Sens. 
Environ., 22, pp. 343-365 
Gonzalez, R. and Woods, R., 1995. Digital Image Processing. 
Prentice Hall, NJ. 
Edwards. K. and Davis .P. A., 1994.The use of Intensity-Hue- 
Saturation transformation for producing color shaded-relief 
images, Photogramm. Eng.Remote Sens, 60(11), pp. 1369-1374 
Mallat. S, 1989. Theory for multi-resolution signal: The 
wavelet representation, IEEE Trans. Pattern Anal. Mach. Intell., 
vol. 11, no. 7, pp.674—693 
Nunez, J., X. Otazu, O. Fors, A.Prades, V, Pala, and R.Arbiol, 
1999.Multiresolution-based image fusion with addtive wavelet 
decomposition, IEEE Transactions on Geoscience and Remote 
Sensing, 37(3): 1204-1211 
Pohl.C and Van Genderen. J. L., 1998. Multi-sensor image 
fusion in remote sensing: Concepts, methods, and applications, 
Int. J. Remote Sens. 19 (5), pp. 823-854.