206 
DISCRETE WAVELET-BASED FUSION 
OF TM MULTI-SPECTRAL IMAGE AND SAR IMAGE DATA 
Liang Shouzhen 3 » Li Lanyong b 
a School of Geomatics, Liaoning Technical University ,Fuxin,liangshzh0816@163.com 
b School of Geomatics, Liaoning Technical University ,Fuxin,lilanyong@sohu.com 
KEY WORDS: Image Fusion, Discrete Wavelet, Multi-spectral, SAR, TM 
ABSTRACT: 
The fusion between optical image data and SAR data is very important. An important reason is that SAR data can complement the 
missing information in optical remote sensing data due to adverse weather condition (rain, cloud cover etc). In this paper, several 
image fusion methods were examined. The experimental results indicate conventional methods including PCA (principle component 
analysis), IHS (Intensity-Hue-Saturation) and Brovey transform are not efficient for fusing TM optical image and SAR image 
because these methods do not take into account the contextual spatial information, while wavelet-based fusion method integrates the 
high-frequency components of the higher resolution data with the low-frequency components of the lower resolution data .So this 
paper adopts the method of 2-D discrete wavelet transform to fuse the TM data and SAR image data. And in this paper a new fusion 
method integrating wavelet and PC is presented. The result shows that the wavelet-based new method improves the high spatial 
texture information apparently and preserves the spectral characteristics of original multi-spectral images highly. Compared with 
conventional fusion methods(HIS,PCA and Brovey transform) ,this method has some advantage according statistic (mean ,entropy, 
standard deviation and correlation ) and better visual result .the output image better preserves the spectral integrity . So it is more 
suitable to fuse TM multi-spectral with SAR image data. 
1. INTRODUCTION 
With the availability of multi-sensor data in many fields, multi 
sensor data fusion has emerged as a new and promising research 
area. The purpose of different sensor data is to better identify 
natural and manmade objects. Namely the essence of image 
fusion is to combine these two kinds of images to form new 
images for improving the performances of the fused images in 
information content, resolution, and reliability and 
interpretation. In the optical remote sensing, some satellite 
sensors supply the spectral bands needed to distinguish features 
but not spatially due to physical and technological constraints, 
while other satellite sensors supply the spatial resolution for 
distinguish the features spatially but not spectrally. Besides the 
optical remote sensing is seriously affected by weather 
condition, for example rain, cloud coverage etc. But the radar 
which is all weather and all time is free of weather condition. 
So in order to improve the data quality, the fusion of multi 
sensor image is very important, especially the fusion of optical 
remote sensing data and radar data. 
Presently many image fusion methods have been proposed for 
combining remote sensing data .A detailed review on this issue 
was given by Pohl and Van Genderen (Pohl et al, 1998).The 
well-known methods are, for example, the HIS (Intensity, Hue, 
Saturation) (Edwards et al, 1994), PCA (Principal components 
Analysis) (Zhou, 1998; Chavez, 1989), Brovey Transform 
(Gillespie et al 1987) .These methods are conventional methods 
which have been studied widely. If the objective of image 
fusion is to construct synthetic images that is closer to the 
reality they represent ,it is no wonder that conventional fusions 
may be used to merge the remote sensing data successfully, for 
example SPOT Pan and multi-spectral data. But these fusion 
methods have some deficiencies. The most serious problem is 
that it leads to some distortion of spectral characteristics. 
Recently developments in wavelet analysis provide a potential 
solution to these drawbacks. For example, Nunez developed an 
approach to fuse a high-resolution panchromatic image with a 
low-resolution multi-spectral image based on wavelet 
decomposition (Nunez et al, 1999). Ranchin and Wald designed 
the ARSIS concept for fusing high spatial and spectral 
resolution images based on the multi-resolution analysis of two- 
band wavelet transformation (Ranchin et al, 2000).Based on the 
virtue of wavelet-based fusion, we adopt this method to merge 
high resolution radar image and low resolution optical image in 
this paper. In addition, we carry on comparison of fusion 
methods according to the criteria proposed by Wald et al (Wald 
et al, 1997). 
The structure of this paper is as follows. The first section is 
introduction .In this section mainly introduce the study on 
image fusion and the structure of the whole paper. The next 
section discusses the theoretical basis and transformation 
characteristics of 2-band wavelet. Then the fusion method based 
on 2-band discrete wavelet is presented. This is followed by a 
discussion of the image fusing experiment. Next, the 
experimental result is compared with previous methods 
developed for image fusion, such as IHS method and PCA 
method. 
2. WAVELET THEORY AND FUSION PROCESS 
Corresponding author iLiang Shouzhen ,School of Geomatics, Liaoning Technical University ,Fuxin, 123000 
E-mail: ,liangshzh0816@ 163 .com