Sersc 3.org (1)

International Journal of Future Generation Communication and Networking
Vol. 13, No. 3, (2020), pp. 3763–3768
3763
ISSN: 2233-7857 IJFGCN
Copyright ⓒ2020 SERSC
Image Fusion of San Francisco Bay SAR images based on ADWT with
optimization technique
J.Thrisul Kumar1
, Dr.Y.Mallikarjuna Reddy2
, M.Venkatesh3
, B.Mani Kanth4
#
1 Asst professor, ECE Dept, VNITSW. e-mail:kumarthrisul9@gmail.com
#2 Principal, Professor in ECE Dept, VVIT, Guntur. e-mail :yennapusa@yahoo.com
#3Asst professor, ECE Dept, VVIT,Guntur.e-mail: venkatesh.munagala@gmail.com
#4 Asst professor, ECE Dept, VVIT, Guntur.e-mail :manikanth439@gmail.com
Corresponding author:kumarthrisul9@gmail.com
Abstract
It is very well-known fact that the monitoring of the Earth is very essential in now a day. Especially
remote sensing is having highest priority for monitoring process. In this regard SAR (Synthetic
Aperture Radar) sensor images are most preferable than optical sensor due to its capturing capability
in any kind of weather conditions. Image fusion is the power tool to extract the required information
from two or more images. In this paper two SAR multitemporal images have been chosen to perform
the fusion. A novel technique ADWT is proposing to perform the image fusion by selecting filter
coefficients through optimization process.Initially DWT is performed with multiple wavelets
(Daubechies, symlet) and it is compared with ADWT (DWT with optimized filter coefficients with BAT
optimization algorithm). Finally, the results are compared in terms PSNR and MSE.
Key words:DWT, ADWT, SAR, BAT, MSE and PSNR
1. INTRODUCTION
Image fusion of the multiple remote sensing images produces the most useful information. Main
aim of merging two or more images is extracting the required content from multiple images. Fusion
could be performed by DWT (Discrete Wavelet Transform). Meaning while two remote sensing
images have been considered for performing fusion [1]. The data images are captured at same
geographical location but at multiple timings. In this regard SAR (Synthetic Aperture Radar) images
are utilized. It is known that various remote sensors(optical sensors and SAR sensors) have been used
to sense the Earth remotely.The reason using SAR sensor other than optical sensor is it can able to
capture the object irrespective of climate conditions.
Employing the ADWT (Adaptive Discrete Wavelet Transform) process rather than
conventional DWT (Discrete Wavelet Transform) is to reduce the error. In any processing system less
error in output is demanded. Consequently, in image fusion process reduction of MSE (Mean Square
Error) and improving the PSNR (Peak Signal to Noise Ratio) is needed. Therefore, in this paper
ADWT is proposing to reduce the MSE which is occurred due to quantization error in DWT process.
In this regard, optimized filter coefficients have been chosen for DWT filter bank instead of
conventional filter coefficients (Daubechies, Symlet). BAT optimization algorithm has been chosen
for selecting the filter coefficients. After selecting the filter coefficients through optimization
algorithm these coefficientsare applied to DWT process which will be called as ADWT [2-7]. Finally,
the results with DWT and ADWT are compared in terms of MSE and PSNR. This paper is organised
as, section II gives the proposing methodology,section III consisting the ADWT, section IV explains
about BAT optimization algorithm and section V gives the results and discussions.

Vol. 13, No. 3, (2020), pp. 3763–3768
3764
II. PROPOSING METHODOLOGY
The proposed methodology of this paper is illustrated in fig.1.
Fig.1. Proposed model
As shown in fig.1. two SAR images are taken for performing the fusion process. In this regard,
conventional DWT has been performed with two types discrete wavelets such as Daubechies and
Symlet wavelet filter coefficients. ADWT process has been implemented by choosing filter
coefficients through BAT optimization algorithm. The output images of DWT and ADWT have been
compared and corresponding MSE and PSNR have been calculated by consideringDWT image with
Daubechies as a reference.
III. ADWT (ADAPTIVE DISCRETE WAVELET TRANSFORM)
Adapting filter coefficients in DWT filter bank through optimization algorithm is known as
Adaptive Discrete Wavelet Transform (ADWT).Consequently, the process of ADWT model has been
depicted in fig.2.Initially BAT optimization process will be performed and corresponding filter
coefficients have been chosen by satisfying thebio-orthonormal property and symmetrical property.
After selection of filter coefficients these will be used for image fusion process.
Fig.2. 2D-ADWT Model

Vol. 13, No. 3, (2020), pp. 3763–3768
3765
As shown in fig.2. single level 2-dimensional DWT and ADWT has been implemented. Main
cause to prefer ADWT rather than conventional DWT is to reduce processing error. During the
process of images in digital type so many stages involved such as ADC, DAC, quantization, encoding
and decoding. In this process MSE might be increased hence in this paper a novel technique ADWT is
proposing to minimize the MSE.
Unfortunately, the reconstructed image from DWT process might not be same as original
image, but there is a chance of minimizing the MSE and PSNR could be improved. Consequently,
losing of the required content in the DWT process could be compensated by adopting optimized filter
coefficients.IDWT (Inverse Discrete Wavelet Transform) will be applied to reconstruct the original
image from the output image of DWT.
IV.BAT OPTIMIZATION ALGORITHM
Fig .2. illustrated the flow chart of BAT optimization algorithm.
Fig.3. BAT algorithm flowchart
BAT algorithmworks according echolocation of the microbats. Microbats emits sonar waves
to identify the prey location it is called as echolocation. BATS can able to see in darkness by using
echolocation of and they can differentiate the prey and enemies. The bats emit loud ultrasonic sound
waves and listen to the echo that reflects back from the surrounding objects. The bat algorithm uses
some idolized rules for simplicity.
(1) Bats use echolocation to sense prey, predator, or any barriers in the path and distance.
(2) Bats fly with a velocity vi and position xi. They have frequency f and loudness ai to reach
their prey. They can adjust the frequency of pulse emission r.

Vol. 13, No. 3, (2020), pp. 3763–3768
3766
(3) As they get close to the prey, pulse increases and loudness decreases.
V. RESULTS AND DISCUSSIONS
Two SAR images have been chosen to verify the proposing technique with conventional
DWT.The data images are the images of San Francisco Bay which are captured by ERS-1 satellite.
Fig.4. a) pre-image b) post-image c) Ground-truth image
Fig.4. represents about the data images which represents the San Francisco Bay.
Fig.5. Fused images a) Daubechies_DWT b) Symlet_DWT c) ADWT_BAT
Fig.5. illustrates the output put images of DWT and ADWT process.Fig.5.a) represents the output
fused image of DWT with Daubechies 2 wavelet coefficients and fig.5.b) represents the output image
of DWT with Symlet wavelet coefficients. Fig.5.c) represents output fusion image of ADWT with
optimized filter coefficients through BAT optimization algorithm.
Measuring
parameter
DWT_Symlet ADWT_BAT
MSE 0.095 0.098
PSNR 30.947 47.767
Table 1. Performance analysis of ADWT and DWT
Table 1 shows the performance of the ADWT with DWT. Fig 6 represents the performance of
proposed methodology graphically.
0
50
100
MSE PSNR
Performamce Analysis
DWT_Symlet ADWT_BAT

Vol. 13, No. 3, (2020), pp. 3763–3768
3767
Fig.6. Graphical representation of performance analysis of ADWT and DWT
VI. CONCLUSION
In this paper, a novel technique ADWT has been proposed for image fusion process. Due to
various processing stages the quantization noise occurs in the conventional DWT. In order to
minimize the noise, the filter coefficients have been adopted through the optimization algorithm. The
proposing ADWT given betterment results for reducing the MSE and to increase PSNR. ADWT is
reduced MSE of 0.30% than conventional DWT with Symlet coefficients. ADWT is better than DWT
with Symlet coefficients of 16.87 for increasing the PSNR.
REFERENCES
[1] Gong, M., Zhou, Z., & Ma, J. (2012). Change detection in synthetic aperture radar images
based on image fusion and fuzzy clustering. IEEE Transactions on Image Processing, 21(4),
2141–2151.
[2] Thrisul Kumar Jakka, Y. Mallikarjuna Reddy, B. Prabhakara Rao “GWDWT-FCM: Change
Detection in SAR Images Using Adaptive Discrete Wavelet Transform with Fuzzy C-Mean
Clustering”, Journal of the Indian Society of Remote Sensing (ISRS), ISSN 0255-660X,
March 2019, Vol 47, No. 3, pp - (379-390).
[3] J. Thrisul Kumar, Y. Mallikarjuna Reddy, B. Prabhakara Rao “WHDA-FCM: Wolf Hunting-
Based Dragonfly with Fuzzy C-Mean Clustering for Change Detection in SAR Images”
published in The Computer Journal (online), on 9th
December 2019. ISSN 0010-4620, EISSN
1460-2067.
[4] J. Thrisul Kumar, Y. Mallikarjuna Reddy, B. Prabhakara Rao “Image Fusion of Remote
Sensing Images using ADWT with ABC Optimization Algorithm” International Journal of
Innovative Technology and Exploring Engineering (IJITEE), ISSN: 2278-3075, Vol -8 Issue-
11, Sep 2019, pp- (3865-3869).
[5] J. Thrisul Kumar, Y. Mallikarjuna Reddy, B. Prabhakara Rao “Change Detection in Sar
images Based on Artificial Bee Colony Optimization with Fuzzy C - Means Clustering”,
International Journal of Recent Technology and Engineering (IJRTE), ISSN: 2277-3878, Vol
-7 Issue-4, Nov 2018, pp- (156-160).
[6] J.Thrisul Kumar, N.Durgarao, E.T.Praveen, M.Kranthi Kumar “ Modified Image Fusion
Technique For Dual-Tree Complex Wavelet Transform” International Journal Of Advanced
Science And Technology , Vol. 29, No. 5s, (2020), pp. 895-901
[7] N.Durgarao, J.Thrisul Kumar , M.Kranthi Kumar, E.T.Praveen “ Novel Dct Co-Efficients
Based Object And Face Recognition” International Journal Of Advanced Science And
Technology , Vol. 29, No. 5s, (2020), pp. 883-888
[8] J. Thrisulkumar, N. Durga Rao, E.T. Praveen, M. Kranti kumar “Image Fusion of Remote
Sensing Kerala Flood Images based on DWT with Multiple Wavelet Families” TEST
Engineering and Management, May – June 2020, ISSN: 0193-4120 Page No. 2345 - 2348
[9] N.Durgarao, J.Thrisul Kumar, M. Kranthi Kumar, E. T. Praveen “ Interactive Segmentation
using improved FLICM” TEST Engineering and Management, May – June 2020, ISSN:
0193-4120 Page No. 2621 – 2624

Vol. 13, No. 3, (2020), pp. 3763–3768
3768
[10] B.M.S. Rani, M.DivyaSree, B. spandana, J.Thrisulkumar “ Novel Retina Recognition
System Using GUI” TEST Engineering and Management, May – June 2020, ISSN: 0193-
4120 Page No. 2316 – 2320
[11] Sweta Srivastava and Sudip Kumar Sahana “Application of Bat Algorithm for
Transport Network Design Problem” Applied Computational Intelligence and Soft
Computing Volume 2019, Article ID 9864090,

Sersc 3.org (1)

More Related Content

What's hot (19)

Similar to Sersc 3.org (1) (20)

Recently uploaded (20)

Sersc 3.org (1)