From the test results and the result tables, we can see that, low 
pass filter is not fit in processing this stripped noise. And the 
other two methods is relatively suiltable. Test result tables show 
that, standard deviation value of processed image by low pass 
filter has much differentce with the proir image and noisy image, 
which suggests this method changes much image information, 
while standard deviation valule of the other two methods 
changes little. 
From the result, we can see that, the brigthness of this processed 
image is reserved and the small stripes like figure 4 decreases. 
Using wavelet transformation of our method, the detail of this 
image can be reserved after noise reduction. 
From these three processed images, we can see that, wavelet 
transformation result has the best visual effect, then gray value 
substition. Though low pass filter can remove noise very well, 
however, some useful information is also removed. In order to 
do a quality comparison, we calculated the mean value, standard 
deviation (std.dev), MSE and PSNR of different results, which 
is seen in table 1. 
1 M-l V-l 
mean = УУ/ (/, j) 
std.dev = 
1 M-\N-l 
YÆ^ mean ~f^j^ 7 
MSE =^7Vr YLmun-giuni 
255" 
where f(i,j) = prior image 
i, j= image coordinates 
M, N = image size 
g(ij) = processed image 
(3) 
(4) 
(5) 
(6) 
MSE 
PSNR 
Noisy image 
Low pass filter 
1260.8 
17.1242 
Gray Value Substitution 
28.0140 
33.6571 
Wavelet transformation 
8.8177 
38.6773 
Table 1. Results and comparisons of MSE and PSNR 
Mean 
Sta. dev 
Prior image 
124.0610 
47.8240 
Noisy image 
153.7807 
53.9107 
Low pass filter 
153.7810 
25.8691 
Gray Value Substitution 
142.4265 
44.0512 
Wavelet transformation 
142.4270 
44.1347 
REFERENCES 
[1] LIU Zheng-jun, WANG Chang-yao, WANG Chen. 
Destriping Imaging Spectrometer Data by an Improved Moment 
Matching Method. JOURNAL OF REMOTE SENSING. 
2002,6(4):279-284. 
[2] Li Yinyan. 2006. Disscussion on Noise Reduction in Image 
Processing.colleget imes.second part of August. 
[3] Garcia J.C., Moreno J.. REMOVAL OF NOISES IN 
CHRIS/PROBA IMAGES: APPLICATION TO THE SPARC 
CAMPAIGN DATA, Proc. of the 2nd CHRIS/Proba Workshop, 
ESA/ESRIN, Frascati, Italy 28-30 April(ESA SP-578,July 2004) 
[4] Dong Guang-xiang, Zhang Ji-xian, Liu Zheng-jun.2006. 
Remote sensing information. 6,pp.36-39. 
[5] Huang Ping S., Su Shun-Chi,and Tu Te-Ming. A Destriping 
and Enhancing Technique for EROS Remote Sensing Imagery. 
JOURNAL OF C.C.I.T., 2004, 32(2) : 
[6] Yang Jinhong, Gu Songshan, Cheng Minghu. 2007. 
Application of Interpolation Method Indestrpping MODIS 
Images. SCIENTIA METEOROLOGICA SINICA. 27(6), pp.605- 
609. 
[7] CHEN Jinsong, ZHU Yuqin, SHAOYun. 2003. Destriping 
Mult-sensor Imagery Based on Wavelet Transform. Remote 
sensing information. 2,pp. 6-9. 
[8] Chang Weiwei. 2007. Study on stripped noise reduction 
methods of hyperspectral image[D], Northwestern Polytechnical 
University. 
[9] Gonzalez Rafael C. 2003. Digital Image Processing (Second 
Edition). Publishing House of Electronics Industry. 
[10] FECIT Sci-Tech.2003. Wavelet Analysis Theory and 
MATLAB Application. Publishing House of Electronics Industry. 
[11] FECIT Sci-Tech.2005 Accessorial Image Processing with 
MATLAB 6.5. Publishing House of Electronics Industry. 
ACKNOWLEDGEMENTS 
The authors wish to thank Chenchao, who is from Shandong 
University of Science and Technology for his software help. 
Table 2. Results of mean and standard deviation value 
216