In: Wagner W„ Szekely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, [APRS, Vol. XXXVIII, Part ?B 
314 
FUSION OF PAN AND MULTISPECTRAL IMAGES BASED ON CONTOURLET 
TRANSFORM 1 
Y. Jia*’ 1 ’, M. Xiao a 
a School of Remote Sensing and Information Engineering, Wuhan University, 129 Luoyu Road, Wuhan 430079,China- 
yhjia2000@sina.com 
b State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan, China - 
yhjia2000@sina.com 
KEY WORDS: Photogrammetry, Fusion, Detection, IKONOS, High resolution, Theory 
ABSTRACT: 
Contourlet transform can give the asymptotic optimal representation of edges and contours, so it can capture smooth contours and 
geometric structures in images much more efficiently than the wavelet transform. The nonsubsampled contourlet transform (NSCT) 
not only has multiresolution and multidirectional properties but also has shift-invariant property by contrast to the contourlet 
transform. A new method of fusing panchromatic (Pan) and multispectral (MS) remote sensing images was proposed based on 
NSCT and IHS transform. IKONOS images were used to verify its validity. The experiment results show that the proposed method 
has greatly improved spatial resolution and keep the spectral information, and is better than the fusion method based on CT and IHS. 
1. INTRODUCTION 
In recent years, the fusion algorithm based on multiscale 
analysis tools is attached a great importance such as 
ridgelet,wavelet, curvelet and contourlet [Arthur L Da C, Zhou 
J P, Do M N, 2006]. As a multiscale analysis tool, the 
contourlet transform (CT) have the characteristics of 
localization, multi-direction and anisotropy[M. Choi, R.Y. Kim, 
M.-R. Nam, H.O. Kim, 2005]. The CT can give the asymptotic 
optimal representation of contours and applied in image fusion 
effectively. However, the CT is lack of shift-invariance and 
results in artifacts along the edges in some extend. The 
nonsubsampled contourlet transform (NSCT) is in virtue of 
nonsubsampled filter banks to meet the shift-invariance [Arthur 
L Da C, Zhou J P, Do M N. 2006]. Therefore, the NSCT is 
more suitable for image fusion. So a fusion algorithm based 
NSCT and HIS (NSCT +HIS) transforms was proposed 
[YANG X H, Jiao L C. 2008]. They demonstrated that the 
NSCT+HIS algorithm was better performance compare to PC A, 
HIS and wavelet transform weighted fusion method. However, 
the high frequent coefficients of the fused image are the sum of 
the high frequent coefficients of intensity and the histogram- 
matched panchromatic image. Different from Yang’s 
method, A new NSCT +HIS fusion method using 
reasonable fusion rule is proposed in this paper, a lower 
resolution MS image of IKONOS and its high resolution 
panchromatic image was fused using the proposed 
method , and the fusion method based on CT and HIS 
(CT+HIS) , simultaneously. The outputs were evaluated in 
objective and subjective performances. Compared with the 
fusion of CT+HIS, the NSCT+HIS fusion technique proved to 
be a better option. 
2. METHODOLOGY 
2.1 Nonsubsampled Contourlet Transform 
Do and Vetterli proposed a ’’true” two-dimensional transform 
called the contourlet transform, which combined with 
nonseparable filter banks and provides an efficient directional 
multiresolution image representation. By virtueof the Laplacian 
pyramids (LP) and directional filter banks(DFB), the CT 
provides the multiresolution decomposition and directional 
decomposition, respectively. The CT cancapture the intrinsic 
geometric structure information of images and achieves better 
expression than discrete wavelet transform (DWT), especially 
for edges and contours. 
Input image 
□ 
Lowpass 
subband 
Bandpass 
subband 
Bandpass 
subband 
Figue 1(a) NSFB structure 
1 The Project Sponsored by the Scientific Research Foundation for the Returned Overseas Chinese Scholars, 
Ministry 
State Education