HYBRID IMAGE FUSION

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 08 | Aug 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1293
HYBRID IMAGE FUSION
Radhakrishna M1, Ullas B C2, Vijay Sakre3, Nikitha T4
1Assistant Professor, Dept. of ECE, Global Academy of Technology, Karnataka, India
2,3,4Student, Dept. of ECE, Global Academy of Technology, Karnataka, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Image fusion is defined as combining useful
information from multiple images to a single image. The
resulting image will be more informative and accurate. Image
fusion not only just combines images but also constructs
images such that the information produced is understandable
from human and machine perception. The satellites capture
images by the high resolution camera and by the help of
concepts of remote sensing. Remote sensing is the process of
gathering the information of an object without making a
physical contact with the object. This paper uses the hybrid
fusion technique which is composed of both spatial and
frequency domains. This paper is based on the concept of
satellite image fusion as an application of hybrid fusion
technique.
Key Words: Remote Sensing, Multisensor, Spatial and
Frequency fusion, Wavelet transform, Fourier
transform.
1. INTRODUCTION
Image fusion combines multisensor data to produce a fused
image with high spatial, spectral, and radiometric
resolutions. Image fusion isthemostbeneficial technologyin
remote sensing for utilizing multisensor,multispectral earth
observation satellites at varying resolutions. Spatial
resolution is critical for delineating objects in a remote
sensing picture. The characteristics of a high spatial
resolution image with multispectral information are easier
to comprehend than a single high resolution Pan image. The
single output image is more informative and accurate than
any of the single source image and it consists ofall necessary
information
Image restoration: Image fusion can be used to restore an
image from morethanonedegradedimageswithuncommon
areas of degradation.
Fig -1: Image Restoration
Image mixing: Two or more images can be fused to create a
new image which carrier more information.
Fig -2: Image Mixing
Image fusion can help in restoration of degradedimagesand
mixing images. By using the hybrid fusion technique, the
image quality will be super enhancedintermsof bothspatial
and frequency domains. Image fusion is used in various
fields like computer vision, remote sensing and medical
imaging.
Image fusion can be broadly classifiedintotwotypes:Spatial
domain fusion and Transform(Frequency)domainfusion.In
spatial domain the operations are done directly on pixels of
the image to get the desired image whereasinthe Transform
domain the operations are done on Fourier transform of the
image followed by Inverse Fourier transform to get the
resultant image.

Fig -3: Fusion
1.1 Spatial Image fusion
Averaging, Select Maximum / Minimum, and Principal
Component Analysis (PCA) are examples of simple image
fusion methods. These methods are classified as spatial
domain methods. The major drawback of spatial domain
approaches is that it produces spatial distortion in fused
images and during the further process spectral distortion
causes negative impact on image causing classification
problems.
1.2IMAGEFUSIONTECHNIQUES(SPATIALDOMAIN)
1.2.1 Simple Average
It is a fusion technique that uses pixel averaging to fuse an
image. This approach focuses on all parts of the image and
works best if the images are captured with the same sort of
photographs sensor. It will provide good results if they have
a high brightness and contrast. This technique is used in
hybrid image fusion for the enhancement of the image.
1.2.2 Minimum Technique
It selects the lowest intensity value of the pixelsfromimages
and produces fused images. This technique is used in
satellite image fusion used in oceanography reflectometry.
1.2.3 Maximum Technique
It selects the pixel values of high intensity from images to
produce fused images. The high intensity pixel is used to
detect the grasslands from the satellite.
1.2.4 Max-Min Technique
It selects the averaging values of the pixels smallest and
largest from the entire source images and produces the
resultant merged image.
2. Frequency Image fusion
In frequency image fusion the Fourier transform of the pixel
is taken and the value of the pixel is achieved by taking the
inverse Fourier transform.
2.1 Wavelet Transform
Wavelets can be defined as the wave-like oscillations
generated according to the frequency of the pixel values.
This transform completely depends on the wavelets of an
image. An image after the wavelet transform application is
divided into four waveletcoefficients,vertical coefficientand
the diagonal coefficient. This procedure keeps onhappening
until the desired image is achieved.
Fig- 4: Image dividing into wavelet coefficients
Both the images are decomposed into wavelet coefficients
using Discrete Wavelet Transform. Only single level discrete
wavelet decomposition of 4 matrices of coefficients.
Fig -5: Four Wavelet Coefficients

Fig -6: Wavelet Coefficients of an image
The given above images are the outputs of the wavelet
transform .The first image is the approximationimageorthe
original image ,the second image is the horizontal
component ,the third is the vertical and the fourth is the
diagonal component respectively. Fusing all the wavelet
coefficients to get the fused wavelets followed bytheinverse
discrete wavelet transform results in the fused image.
Fig -7: Fusion Process
The wavelet coefficients of the two images can be fused
using different combinations of mathematical operations.
Fusion is performed in two ways
1. fusion1 for approx. coefficient
2. fusion2 for detailed coefficient
LL=fusion1 (LL1, LL2)
HL=fusion2 (HL1, HL2)
LH=fusion2 (LH1, LH2)
HH=fusion2 (HH1, HH2)
Fusion1 and fusion2 are the mathematical operations such
as mean, max and min.
Therefore total 9 combinations of mathematical operations
on approximation coefficients and detailed coefficients are
possible such as
MeanMean, MeanMax, MeanMin, MaxMean, MaxMax,
MaxMin, MinMean, MinMax, MinMin
Table -1: Mathematical operations
Based on the data acquisition of a user and by observing the
efficiency of the algorithm, a suitable method can be chosen
for the appropriate result.
Fig -8: MaxMax
Fig -9: MinMax
Fig -10: MeanMax

Nowadays satellite communication is being advanced at a
higher level using the concept of remote sensing, where the
information of the image is acquired without the physical
contact of the object. In the hybrid fusion technique spatial
and frequency domains are combined for the enhancement
of the image.
3.1 Working of Hybrid Image Fusion
The hybrid image fusion uses the concept of bothspatial and
frequency domain. The algorithm processes two imagesand
fuses both in spatial domain and frequency domain and
takes the sum of products of results of fusion of the first
image and the second image with the given value alpha and
beta respectively.
The algorithm multiplies 0.5 to the fusion values (fusion1
and fusion2) and takes the sum of it to get the final hybrid
fusion value
The algorithm takes the sum of products of fusion1 and
fusion2 with 0.5 to get the final fusion value.
The algorithm takes the sum half of fusion1 and fusion2 to
get the final value of fusion to get the final fusion value.
fusion = alpha * fusion1+ beta * fusion2
Fig -11: Block diagram Hybrid Image Fusion
Fig -12: Fused Output
4. CONCLUSION
By using this method of the Hybrid image fusion we can
achieve the better quality image by the combination of both
spatial and the frequency domain. The resultedimagecan be
used as application for the military and the experimental
purpose.
REFERENCES
[1] TWO STAGE SPATIAL DOMAIN IMAGE FUSION
TECHNIQUES C. Morris and R.S. Rajesh ISSN: 0976-
9102M. Young, The Technical Writer’s Handbook. Mill
Valley, CA: University Science, 1989.
[2] World Academy of Science, EngineeringandTechnology
International Journal of Geological and Environmental
Engineering Vol: 11, No: 9, 2017.
[3] Image Fusion Techniques-A Comparative Study Vibha
Gupta, Sakshi Mehra,International Journal of
Engineering Trends and Technology (IJETT) – Volume
32 Number 2- February 2016.
[4] Image Fusion Based on Wavelet Transformation
Raghawendra Bhimarao Naik, Pavan N.Kunchur
International Journal of Engineering and Advanced
Technology (IJEAT) ISSN: 2249 – 8958 (Online),
Volume-9 Issue-5, June 2020.
[5] Satellite Image Fusion using Fast Discrete Curvelet
Transforms C.V.Rao, J.Malleswara Rao, A.Senthil Kumar,
D.S.Jain, V.K.Dadhwal National Remote Sensing Centre,
Indian Space Research Organization, Hyderabad-
500037, India.
[6] Image Fusion Techniques: A Survey Harpreet Kaur,
Deepika Koundal & Virender Kadyan Archives of
Computational Methods in Engineering Volume 28,
4425–4447 (2021).
[7] A Hybrid Image Fusion Algorithm for Medical
Applications Written by Appari Geetha Devi, Surya
Prasada Rao Borra and Kalapala Vidya Sagar Submitted:
September 30th, 2020 Reviewed: March 2nd, 2021
Published: April 13th, 2021.
3. Hybrid Image Fusion

HYBRID IMAGE FUSION

More Related Content

Similar to HYBRID IMAGE FUSION (20)

More from IRJET Journal (20)

Recently uploaded (20)

HYBRID IMAGE FUSION