You are on page 1of 5

(IJCSIS) International Journal of Computer Science and Information Security,

Vol. 10, No. 1, January 2012


Denoising Cloud Interference on Landsat Satellite
Image Using Discrete Haar Wavelet Transformation

Candra Dewi
Department of Mathematic
University of Brawijaya
Malang, Indonesia
d3w1_c4ndr4@yahoo.com
Mega Satya Ciptaningrum
Department of Mathematic
University of Brawijaya
Malang, Indonesia
megasatya@yahoo.com
Muh Arif Rahman
Department of Mathematic
University of Brawijaya
Malang, Indonesia
arifrahman@ub.ac.id



AbstractSatellite imagery is very useful in information
acquisition of the earth's surface image, especially the earth's
resources. However, in the process of retrieval information from
satellite imagery is often found barriers that can obscure or even
cover the imaging of an area. One of these barriers is a cloud,
which result the image that covered with lots of noise. Wavelet
transformation was usually used to enhance the image or to
eliminate striping noise on satellite image. In this paper is used
Discrete Haar Wavelet transformation to reduce cloud noise on
Landsat TM image. The process includes the Haar Wavelet
decomposition of image rows and columns. After that,
thresholding process is also applied for de-noising. Thresholding
results are then reconstructed using the Inverse Discrete Haar
Wavelet. The method is applied to the variation of the band
image, the type of thresholding (hard and soft), as well as the size
of the image convolution. The testing results on the band 1 to
band 6 of Landsat TM imagery showed that the lowest error
values are calculated by RMSE (Root Mean Square Error)
present in band 1. Image signal to noise ratio in band 1 has the
highest value, which means most high-power image signal to
noise. This mean that band 1 has the highest pixel value
similarity between whole testing data.
Keywords; Discrete Haar Wavelet, thresholding, image
convolution, Landsat TM

I. INTRODUCTION
Image of the earth surface recording can be interpreted by
the user for the benefit of various fields. In the process of
image acquisition by the satellite, sometimes is found noise that
can reduce the image quality. This disorder is caused by the
presence of such clouds or fog that can obscure or even
covered the satellite during the imaging process [1]. This noise
can interfere the interpretation process therefore the results
obtained will not be maximal.
Each pixel in the satellite image has some digital value
(numeric) in accordance with the band of satellite imagery. For
example is Landsat TM image that has 7 bands. Therefore,
each pixel has 7 digital values that are suited to 7 band digital
value that is owned. The different characteristic each bands
causes the difference in the ability to detect clouds. In the study
that is performed by Choi and Bindschadler (2004), clouds is
very high reflected in the band 2 (0.52 - 0.60 m).
The elimination process of noise in the spatial domain can
be applied directly on image pixels. One of the transformation
methods that can be done on the spatial domain is a power-law
transformation. While in the frequency domain, the image is
broken into multiple kernels to be processed by the analysis of
transformation. Transformations that can be done in this
domain include Wavelet transformation [3] [4] [5].
Transformation performed to obtain information and identify
the original image, by getting its spectrum. Spectrum can be
obtained from the image frequency, time, or time-frequency
depend on the type of transformation used [6].
It is well known that wavelet transform is a signal
processing technique which can display the signals on in both
time and frequency domain. Wavelet transform is superior
approach to other time-frequency analysis tools because its
time scale width of the window can be stretched to match the
original signal, especially in image processing studies.
Wavelet transformation can be used to obtain signal both in
the frequency domain and time domain. Wavelet time scale
width of the window can be stretched to match the original
signal. Wavelet is a conversion function that can be used to
break up a function or a signal into different frequency
components. These components then can be processed in
accordance with the scale. While the wave is a function of
moving up and down in space and time periodically
(sinusoidal), wavelet is a limited wave or sometimes is called
as short wave [7].
Haar transform uses the Haar scaling function and the Haar
wavelet function. Haar wavelet transformation use the Haar
basis functions that is called a wavelet orthonormal [8]. Haar
Wavelet functions can be expressed in matrix form.
In the previous study, wavelet transform is used to sharpen
the cloud-related shadow areas [1]. Beaulieu et al (2003) refine
the resolution of a multi-spectral (MS) image using fusion
method and the Stationary Wavelet Transform. In the study
performed by Torres and Infante (2001), wavelet transform is
used for denoising stripping noise on satellite imagery. This
27 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 10, No. 1, January 2012
paper applies the Haar wavelet transformation to reduce the
noise cloud on Landsat TM imagery.
II. PREVIOUS RESEARCH
The research about the using of wavelet transformation has
been done by some researcher. Torres and Infante (2001)
present new destriping technique on satellite imagery using
Daubechies wavelets of different orders and was tested on a
heavily striped Landsat MSS image. Visual inspection and
measurement the signal-to-noise ratio showed that the method
proved produce encouraging results in image quality and
performance, overcome some problems commonly found on
traditional destriping techniques and reduce computer time
process and storage space.
Beaulieu et al, (2003) refine the resolution of a multi-
spectral (MS) image by fusion method using a high-resolution
panchromatic (PAN) image and the Stationary Wavelet
Transform (SWT). They propose to produce high-resolution
MS image that has nearly the same statistical properties than
the original multi-spectral image with no blocking image
artifacts. These algorithms are based on the injection of high-
frequency components from the PAN image into the MS
image. They prove that pixel-level fusion was a powerful
method to refine the spatial resolution of PAN images.
Wang et al (2003) present a new approach to eliminate the
random image noises inherent in the microarray image
processing procedure using stationary wavelet transform
(SWT) and applied on analysis of gene expression. The testing
result on sample microarray images has shown an enhanced
image quality. The results also show that it has a superior
performance than conventional discrete wavelet transform and
widely used adaptive Wiener filter in this procedure.
Elrahman and Elhabiby (2008) developed image sharpening
algorithm using wavelet to enhance shadow areas of cloud and
tested this algorithm on the panchromatic band of Landsat 7
ETM satellite sub-scenes. The algorithm is applied locally by
boosting the image high frequency content in the shadow areas
using the defected image de-noised wavelet coefficients. By
using visual and quantitative analysis was found that the ability
to enhance details under shadow areas increased with the
increase in the number of wavelet decomposition levels.
Beside, were found that enhancing image quality in the shadow
areas could be done using only two or three wavelet
decomposition levels.
In these previous studies, the using of wavelets on the
satellite image is to sharpen the image and to improve image
resolution. Wang et al (2003) already used wavelet to eliminate
the noise, but is applied to the gene sequence image. In this
paper will be applied discrete Haar wavelet to reduce noise in
the form of clouds on satellite images. Although the discrete
wavelet transform has a lower performance of the stationary
wavelet transform, but its ability to reduce the noise is quite
high and does not vary with stationary wavelet transform [5].
III. RESEARCH METHOD
This application was built to reduce noise on Landsat TM
satellite image using Haar wavelet transformation method. The
limitation of this system includes:
1) The image used is a grayscale image of type TIFF
2) The size of the image used is 256x256 orthonormal

The flowchart of noise reduction process is shown Figure 1.
The inputs of this application consist of satellite imagery with
clouds noise and image without noise. This input image is
presented in grayscale values. Some preprocessing was done to
the noise image to reduce the noise. The image without noise is
used as a comparison in the testing process.














Figure 1 Flowchart of noise reduction process

Firstly, noise image is transformed into the frequency
domain using the Haar wavelet transform. The quantization
process is then performed using a specific threshold value. The
transformation process is performed to the n level, where N =
2
n
and N is the size of the image. At each level, the row
transformation is done in advance through highpass and
lowpass filters. After that, is done transformation of the
column.
The next process is tresholding. This process separates
pixels based on the degree of grey level values. The wavelet
coefficients which are below the threshold are set to zero and
than take the other values for purposes of reconstruction of the
signal. Threshold used is Hard and Soft Threshold. With is
the threshold value, hard thresh equation is shown in (1).

>
=

| | 0
| | ,
) (
x
x x
X T
hard
(1)
On the hard threshold, all wavelet coefficients with a
value below a specified threshold are classified as noise and
removed (are set to zero). While the coefficients above the

28 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 10, No. 1, January 2012
threshold is classified as signal. In soft thresholding, the
wavelet coefficients with a value below the specified threshold
are removed and the wavelet coefficients above the specified
threshold are reduced by the threshold value. Thus, this method
reduces the range of wavelet coefficients and signal leveling.
Soft threshold was chosen because this procedure does not
cause non-continuants at x = . The equation for the soft
threshold is shown in (2).

<

=

| | , 0
| | ), | )(| (
) (
x
x thresh x x sign
X T
soft
(2)
For the determination of threshold values is used equation
as in (3).
n
n
t
) log( 2
2

=
(3)
Where:
t = threshold value that is calculated
= the variance of data
n = number of data

The equation of variance is shown in (4).
) 1 (
) (
2
2

=

n
x x
i

(4)
The last process is the Inverse Haar Wavelet Transform
(IHWT) which is the process of passing the image through the
inverse filter matrix transformations. This process is contrary to
the decomposition process.
IV. TESTING METHOD

For testing the result is used Root Mean Square Error
(RMSE) and Peak-to-Signal Noise Ratio (PNSR).
A. Root Mean Square Error (RMSE)
RMSE is one of the ways to measure the amount of the
difference between the estimated values with actual values by
measuring the average of error. RMSE is calculated by
comparing the number of errors between the denoising image
and the original image. The lower the RMSE value the smaller
the error calculation has been done. RMSE of digital image
with size NxM could be calculated using equation as shown in
(5).
2
2
)] , ( ) , ( [
N
j i F j i f
RMSE


=
(5)
Where:
f(i,j) is pixel value in original image
F(i,j) is the pixel value on reconstruction image
N
2
is an image size (in pixels)


B. Peak-to-Signal Noise Ratio (PNSR)
PSNR is the comparison between the maximum possible
signal strength of a digital signal with the power of noise that
affects on the signal (Alfatwa, 2005). PSNR is defined through
the signal-to-noise ratio (SNR) to measure the level of signal
quality. Signal quality is directly proportional to the value of
SNR. The larger of the SNR value, the better the quality of the
generated signal. PSNR values usually range between 20 and
40 dB (Alfatwa, 2005). PSNR values can be calculated using
equation as shown in (6). Value of 255 represents the upper
limit value of image pixels.
|

\
|
=
RMSE
PSNR
255
log 20
10
(6)

V. SOURCE OF DATA

Image that is used in the testing process is Landsat TM
satellite image with each channel has a different sensitivity to
the wavelength. Landsat TM satellite orbital period for taking
pictures of the earth's surface is generally performed at least 6
months. Satellite imagery from two period of taken picture can
be used as a reference on the interpretation process. For
example, this study used two images with the same object (the
island of Madura) taken in June 2004 and February 2005. In
the image taken on 2005 exists cloud covering the particular
object and the image taken on 2004 (with the same object) is
used as reference.
Preparation of satellite imagery should be done to obtain
the image that is suited to analysis. The original image is
cropped to the size of 256 x 256 pixels and converted into Tif
extention format. In addition, the original image with 7 bands is
separated per-band for used in applications. Details of the data
used are as follows:
1) Landsat image of Madura island, dated June 25, 2004
and dated February 4, 2005
2) Landsat image of Java island, dated June 25, 2004 and
dated February 4, 2005

Of the two sources of image data was made 2 pieces of
testing data with each of the data contained six band image data
(bands 1 to 6 / 7) with each size is 256 x 256 pixels.
Data I: latitude 7:7:45.66 S and longitude 113:3:12. 43 E
Data II: latitude 7:39:41.99 S and longitude 112:56:41.87 E
VI. RESULT AND DISCUSSION

Some examples of images resulted from denoising process
are visually displayed in Table I. The first image shows the
result of denoising on data I (band 1) with convolution 2 (hard
thresholding), the second on data II (band 1) with convolution
8 (soft thresholding), and the third on data II (band 3) with
convolution 8 (hard thresholding).
29 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 10, No. 1, January 2012
TABLE I. SAMPLE IMAGE OF TESTING RESULT
No Band Input Images Output Images

1
1
(data I)

2
1
(data II)

3
3
(data II)


The RMSE was calculated in the image (the data I and II)
which has been transformed with Haar Wavelet. This RMSE
values are calculated against several variations of testing which
includes testing of inter-thresholding methods, inter-level
convolution, and inter-band image. Furthermore, the RMSE is
used as input to the calculation of PNSR to observe the ratio of
signal strength to noise.
Based on the results in Table 1 could be known that
visually processes of Haar wavelet denoising did not show the
significant results, because the cloud noise in each band is
represented differently. Therefore, an analysis on the basis of
testing results on PNSR and RMSE are performed.
The comparison results of RMSE and PNSR on bands 1 to
7 with a convolution of 8, 4, and 2 are shown in Table 2 to
Table 7. The RMSE and PNSR are obtained can be used to find
out the best band on Landsat satellite imagery for the cloud
denoising process.
TABLE II. STHE CALCULATION RESULT OF THE RMSE AND PNSR AT
CONVOLUTION 8 (DATA I)
RMSE PNSR
Band
Thresh
Value
Hard
Threshold
Soft
Threshold
Hard
Threshold
Soft
Threshold
1 3,34 30,128 29,633 18,551 18,695
2 2,93 30,98 30,356 18,309 18,486
3 2,93 38,246 37,681 16,479 16,608
4 2,78 42,456 41,998 15,572 15,666
5 2,77 46,93 46,548 14,702 14,773
6/7 2,12 44,859 44,569 15,094 15,15
TABLE III. THE CALCULATION RESULT OF THE RMSE AND PNSR AT
CONVOLUTION 8 (DATA II)
TABLE IV. THE CALCULATION RESULT OF THE RMSE AND PNSR AT
CONVOLUTION 4 (DATA I)
RMSE PNSR
Band
Thresh
Value
Hard
Threshold
Soft
Threshold
Hard
Threshold
Soft
Threshold
1 3,27 31,306 30,463 18,274 18,455
2 2,86 32,353 31,691 17,932 18,112
3 2,86 39,3 38,646 15,243 16,389
4 2,72 43,3 42,702 15,401 15,522
5 2,71 48,174 47,818 14,475 14,539
6/7 2,08 46,025 45,744 14,871 14,924
TABLE V. THE CALCULATION RESULT OF THE RMSE AND PNSR AT
CONVOLUTION 4 (DATA II)
RMSE PNSR
Band
Thresh
Value
Hard
Threshold
Soft
Threshold
Hard
Threshold
Soft
Threshold
1 2,93 31,988 31,1 18,031 18,276
2 2,57 33,509 32,773 17,628 17,82
3 2,57 40,198 39,393 16,047 16,222
4 2,45 43,854 43,025 15,291 15,456
5 2,45 49,071 48,754 14,314 14,371
6/7 1,89 46,746 46,469 14,736 14,788
TABLE VI. THE CALCULATION RESULT OF THE RMSE AND PNSR AT
CONVOLUTION 2 (DATA I)
RMSE PNSR
Band
Thresh
Value
Hard
Threshold
Soft
Threshold
Hard
Threshold
Soft
Threshold
1 2,89 7,647 7,362 30,461 30,791
2 2,4 10,745 10,262 27,507 27,906
3 2,01 20,123 20,141 22,057 22,049
4 2,86 43,812 43,018 15,299 15,458
5 2,62 26,025 25,486 19,823 20,005
6/7 1,57 22,894 22,873 20,936 20,944
RMSE PNSR
Band
Thresh
Value
Hard
Threshold
Soft
Threshold
Hard
Threshold
Soft
Threshold
1 2,96 7,422 7,041 30,72 31,178
2 2,46 10,339 9,863 27,841 28,251
3 2,06 19,748 19,069 22,22 22,282
4 2,92 43,214 42,623 15,418 15,538
5 2,69 25,312 24,773 20,064 20,251
6/7 1,61 22,551 22,421 21,057 21,118
30 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 10, No. 1, January 2012
TABLE VII. THE CALCULATION RESULT OF THE RMSE AND PNSR AT
CONVOLUTION 2 (DATA II)
RMSE PNSR
Band
Thresh
Value
Hard
Threshold
Soft
Threshold
Hard
Threshold
Soft
Threshold
1 2,59 7,927 7,835 30,148 30,25
2 2,15 11,139 10,667 27,194 27,57
3 1,8 30,564 20,864 21,869 21,75
4 2,56 44,428 43,333 15,178 15,394
5 2,35 26,787 26,279 19,572 19,739
6/7 1,41 23,313 23,478 20,779 20,178

Limitations of different threshold values applied to each
image because the distribution of each image pixels values is
different. This research calculates the threshold based on the
characteristics of image to obtain the best threshold value.
From Table 2, Table 4 and Table 6 can be seen that the
band 1 has the smallest RMSE values both in hard thresholding
and soft thresholding method (both in the convolution 8, 4, and
2). The lowest RMSE values observed in the convolution 8
with soft thresholding, which is about 29.633 (Data I) and
7.041 (data II). Since the highest RMSE value is detected on
band 5 (using hard thresholding with convolution 2) that is
about 49.071 (Data I) and 26.787 (Data II). The quite far
differences of RMSE value is caused by variations in image
value. Data I is an image with a lot of noise distribution, while
the data II has less noise in the form of clouds. Base on RMSE
value can be seen that band 2 has the lowest error values and
band 5 has the highest error value.
The highest value of PNSR is observed on the band I that is
around 18.695 dB (data I) and 31.178 dB (data II), while the
lowest value is observed in the band 5 with the value is 14.314
dB (data I) and 19.572 dB (data II). It means that the ratio of
the image signal to noise at a band I higher than the band 5.
The Signal strength value at Data II tends to be higher than the
data I, because the noise in the form of clouds fewer than on
the Data I. It denoted that the highest probability to perform
denoising of cloud can be done on the band 1. On the contrary
the lowest probability is on band 5. These results are quite
relevant to the characteristic of a band I with a wavelength of
0.45 to 0.52 m which serves to increase penetration on water
body and humidity.
VII. SUMMARY AND CONCLUDING REMARKS

In this paper, Discrete Haar Wavelet methods is applied by
utilize thresholding method to the data testing (Landsat satellite
image with size 256x256 pixels) to reduce the noise contained
in image. The testing result shows that the lowest RMSE value
is detected on band 1 (29.633) and highest value is on the band
5 (49.071). As well as the highest PNSR value observe on the
band I (18.695 dB) and lowest value is on band 5 (14.314 dB).
It can be concluded that the best band to perform denoising
clouds with Haar Discrete Wavelet found on the band I, and
worst band found on the band 5.
For further study, is proposed to test the result referable
reinforced with a system of classification on Landsat satellite
imagery.

REFERENCES

[1] A. Abd-Elrahman and M. Elhabiby, Wavelet Enhancement of Cloud-
Related Shadow Areas in Single Landsat Satellite Imagery, Beijing:
The International Archives of the Photogrammetry, Remote Sensing, and
Spatial Information Science, Vol. XXXVII part B7, p.1247-1252, 2008.
[2] H. Choi dan R. Bindschadler, Cloud Detection in Landsat Imagery of
Ice Sheets Using Shadow Matching Technique and Automatic
Normalized difference Snow Index Threshold Value Decision, Remote
Sensing of Environment, Vol. 91. p.237-242, 2004.
[3] J. Torres and S.O. Infante, Wavelet Analysis for The Elimination of
Striping Noise In Satellite Images, Society of Photo-Optical
Instrumentation Engineers, DOI: 10.1117/1.1383996, 2001.
[4] M. Beaulieu, M., S. Faucher, and L. Gagnon, Multi-Spectral Image
Resolution Refinement Using Stationary Wavelet Transform,
International Geoscience Remote Sensing Symposium, Vol. 6, pp. 4032
4034, 2003.
[5] X.H. Wang, Robert S.H. Istepanian, and Y.H. Song, Microarray Image
Enhancement By Denoising Using Stationary Wavelet Transform,
IEEE Transactions on Nanobioscience, Vol. 2, No.4, 2003.
[6] D. F. Alfatwa, Watermarking pada Citra Digital Menggunakan Discrete
Wavelet Transform, Informatic Study Program, Technoly Institute of
Bandung, 2005.
[7] R. B. Edy Wibowo, Scattering Problem for A System of Non Linear
Klein-Gordon Equations Related to Dirac-Klein-Gordon Equations, An
International Multidisciplinary Journal, Vol. 71, No. 3-4, 2009.
[8] Gonzales, Rafael C dan Woods, Richard E. 2005. Digital Image
Processing, 2
nd
Edition. New Jersey: Prentice Hall



31 http://sites.google.com/site/ijcsis/
ISSN 1947-5500