The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part Bl. Beijing 2008 
(c) (d) 
Figure5. Result of selected filters on another SAR image 
(a) Lee, (b) Gamma-MAP, (c) Wiener, (d) ISPW 
The results of this experiment show that ISPW method achieved 
similar results as that of experiment I except RR values. 
Denoised image of ISPW method gives the best visual quality 
as well as the highest SNR, PSNR and EPA values in 
comparison of the rest filters. However, ENL is still not 
optimistic in this experiment, and RR cannot retain a good 
result as foregoing experiment. According to definition of 
radiation resolution, it is proportional to ratio of deviation to 
mean, therefore, it can be easily compensated by grey linear or 
non-linear transform. 
4. CONCLUSIONS 
In this paper, a new technique to reduce speckle noise in SAR 
images has been proposed. On the basis that pixels of log- 
transformed SAR images are mutually independent, this insitu 
single-pointed wavelet-based approach makes use of 
approximate component information of wavelet coefficient in 
each scale to deal with approximate and detail part in 
corresponding scale in order to suppress speckle noise locally 
and not affect other useful information. A generally 
comprehensive series of evaluation criteria is applied to analyze 
the performance of this method. Experiment results show that 
this method gives a comparatively better image with more 
detailed edge information and more clean objects. 
RR 
ENL 
SNR 
PSNR 
EPA 
Original 
1.5015 
1.3793 
— 
— 
— 
Lee 
1.2056 
2.2287 
14.1288 
21.1549 
0.3904 
Gamma-MAP 
1.1930 
2.2831 
13.5733 
20.5993 
0.3690 
Wiener 
1.2447 
2.1147 
14.1557 
21.1817 
0.4325 
ISPW 
1.3307 
1.6650 
16.0485 
23.0745 
0.6667 
Table4. Statistical result of selected filters on another SAR image 
ACKNOWLEDGEMENTS 
During the research we gain lots of help from others. Thanks 
are due for the support from the Natural Science Fund of P. R. 
China (No. 40601084 and No. 40523005), the Open Research 
Fund Program of State Key Laboratory of Satellite Ocean 
Environment Dynamics (No. SOED0602), the Open Research 
Fund Program of the Geomatics and Applications Laboratory 
of Liaoning Technical University, Open Research Subject of 
Key Laboratory of Geo-informatics of State Bureau of 
Surveying and Mapping (No. A1721), China International 
Science and Technology Cooperation Project: High-Resolution 
Stereo Mapping Satellite Field Geometric Calibration and 
Application (No. 2006DFA71570), Commission of Science 
Technology and Industry for National Defense Project: Key 
Techniques of Data Processing for Mapping Satellite and 
China National 863 Project: Intensive Image Matching and 
Continuous Digital Surface Reconstruction Method Based on 
High Overlap Aerial Images (No. 2006AA12Z134). With those 
help, our research is able to go along propitious. 
REFERENCES 
Anurat Ingun, 1996. Reducing speckle noise in SAR imagery 
using wavelet transforms and high-order statistics. Dissertation 
of Florida Institute of Technology, chapter 4, pp. 66-69. 
Chang S G, Yu B and Vetterli M, 2000. Adaptive wavelet 
threshold for image denoising and compression. IEEE 
Transaction on Image Processing, vol. 9, no. 9, pp. 1532-1546. 
Donoho D L and Johnstone I M, 1994. Ideal spatial adaptation 
via wavelet shrinkage. Biometrika, vol. 81, no. 3, pp. 425-455. 
Frost, V. S. et al, 1982. A model for radar images and its 
application to adaptive digital filtering of multiplicative noise 
[J]. IEEE Transaction on Pattern Analysis and Machine 
Intelligence, vol. 4, no. 5, pp. 157-166. 
Jansen M, Malfait M and Bultheel A, 1997. Generalized cross 
validation for wavelet threshold. Signal Processing, vol. 56, no. 
l,pp. 33-44. 
J. W. Goodman, 1976. Some fundamental properties of speckle, 
J. Opt. Soc. Amer., vol.66, pp. 1145-1150. 
Lee, J. S, 1980. Digital image enhancement and noise filtering 
by use of local statistics [J], IEEE Transaction on Pattern 
Analysis and Machine Intelligence, vol. 2, no. 2, pp. 165-168.