Digital Image

Fusion
 What is Image Fusion ?
Combining different images to form an information rich high-resolution image

Panchromatic Multi-Spectral  Fused Image

 Why is Image Fusion Needed ?

Complementary and redundant information from imaging sensors

• Eg: Visible image provides the outline of scene, while infrared image can show the existence of some special
objects, such as concealed guns or people

Limitations of available data for our needs

• Eg: Satellites such as SPOT, Landsat, IKONOS etc provides panchromatic images at high spatial resolution and
multispectral images at low spatial resolution.But many remote sensing applications require both high spatial and
high spectral resolutions

Presence of noise and focus errors in images

• Eg: Getting clear image of a fast-moving car using traffic camera

Other needs

• Image sharpening, Improve geometric corrections, Improved classification, Detect changes using multi-temporal
data, Replace defective data etc
 Steps in Image Fusion

Image registration Decision Labeling  Radiometric calibration

Align the subsequent features of Operator labels the registered Employed on spatially aligned
various images with respect to a images with respect to pixel or images.Transformation of feature
reference image. feature maps. A set of decision maps is performed on an
maps are produced ordinary scale to get result in a
similar representation format. 

Obtains the decision or feature Merges the consequential images

Significant features of registered maps that might not pass on to a into a one resultant image
images are extracted to produce similar object.Connects maps to a containing an enhanced
several feature maps. common object to perform fusion explanation of the image

Feature extraction Semantic equivalence Image Fusion

 Image Fusion Categories

MULTIVIEW MULTIMODAL MULTITEMPORAL MULTI FOCUS FUSION FOR

FUSION  FUSION  FUSION  FUSION  IMAGE
RESTORATION
Image Fusion
Categories Contd...

Multiview Fusion
• Images of the same modality,
taken at the same time but from
different places or under
different conditions
• Goal: to supply complementary
information from different
views
Image Fusion
Categories Contd...

Multimodal Fusion
• Images of different modalities:
PET, CT, MRI, visible, infrared,
ultraviolet, etc.
• Goal: to decrease the amount of
data, to emphasize band-specific
information 
Image Fusion
Categories Contd...

Multitemporal Fusion
• Images of the same scene taken
at different times (usually of the
same modality) 
• Goal: Detection of changes 
Image Fusion
Categories Contd...

Multifocal fusion 
• The original image can be
divided into regions such that
every region is in focus in at
least one channel 
• Goal: Image everywhere in
focus
Image Fusion
Categories Contd...

Fusion for image restoration 

• Idea: Each image consists of
“true” part and “degradation”,
which can be removed by fusion
• Types
• Image denoising
• Blind deconvolution
• Super resolution
Image Fusion
Techniques
• In spatial domain, we deal with
images as it is. The value of the
pixels of the image change with
respect to scene
• In frequency domain, we deal
with the rate at which the pixel
values are changing in spatial
domain.
 Simple Average

• Combines images by averaging the pixels focused on all regions of

the image. Used for high brightness and high contrast images

Minimum Technique

• Selects lowest intensity value of the pixels from images. It is used

for darker images
Image Maximum Technique
Fusion Techniques
• Selects pixels values of high intensity from images .It is used for
Contd... lighter images
(Spatial Domain)
Max–Min Technique

• Selects averaging values of the pixels smallest and largest from the
entire source images and produced the resultant merged image.

Simple Block Replace Technique

• Adds all images of pixel values and takes the block average for it. It
is based on pixel neighboring block images.
 Weighted Averaging Technique

• Assigns weights to every pixel in source images and resultant image produced
by the weighted sum of pixel values. Used to improve detection reliability of the
output image

Hue Intensity Saturation (HIS)

• Converts RGB image into HIS components and intensities divided

with panchromatic (PAN) image followed by inverse transformation. Provides
best result for remote sensing images
Image Brovey Transform Method
Fusion Techniques • Standardizes RGB bands to append the intensity and brightness into the image .
Contd... Includes an RGB color transform technique and  is helpful for visual
(Spatial Domain) Interpretation but generates spectral distortion 

Principal Component Analysis (PCA)

• Orthogonal transformation for converting observations with correlated variable

into principal components. Causes spectral degradation and color distortion.

Guided filtering

• Boundary smoothing  preserving operator having enhanced performance near

boundaries.Provides better results as compared to Principal component analysis 
 Laplacian Pyramid Fusion Technique

• Uses the interpolation sequence and Gaussian pyramid for multi-

resolution analysis for image fusion. 

Discrete Transform Fusion Method

• Components of the images are separated subsequently; discrete

transformation is applied, and the average of  images computed an
Image inverse transformation is applied at the end to obtain a fused image.

Fusion Techniques Discrete Cosine Transform (DCT)

Contd... • DCT done based on DCT parameters of magnitude,contrast
(Frequency Domain)
measure, contrast highest,Energy etc.Used in factual time applications

Discrete Wavelet Transform (DWT) Method

• Decomposes Images into various high and low-frequency bands,

minimizing the spectral distortion in the resultant images,producing good
signal to noise ratio.Superior to the spatial domain in minimizing color
distortions
 Kekre’s Wavelet Transform (KWT) Method

• Generates KWT matrices from Kekre’s transform method matrix . Used

for more than one images, and the fused image is far good than other
methods.

Kekre’s Hybrid Wavelet Transform (KHWT) Method

• KHWT method has been derived from hybrid wavelet transforms. In this

Image method, the best two matrices are combined into a hybrid wavelet
transforms method
Fusion Techniques
Contd... Stationary Wavelet Transform (SWT) Method

(Frequency Domain) • New wavelet transform method with translation invariance to overcome

disadvantages of DWT method. It provides enhanced analysis of image
facts

Curvelet Transform Method

•  Curvelet is a new multi-scale transform; it breaks the disadvantages of

wavelet method in representing directions of boundaries in the image
 Remote Sensing Applications

• Temporal Image fusion used to combine Landsat and moderate resolution imaging
spectroradiometer data
• Image Fusion provides high detail data during Disasters
• Overcoming image limitations in spatial and spectral resolution

Medical Domain Applications

Applications • CT and MRI used collectively with IF techniques to enhance accuracy and sensible applicability
• Fusing of Medical imaging data for non-invasive diagnosis

Surveillance Domain Applications

• Fusion of visible and infrared images has been introduced for face recognition, image dehazing,
and military reconnaissance etc

Other Applications

• Object recognition, tracking, object detection, Crime Investigation etc

Image Fusion Tools in ArcGIS Pro
& ENVI

 ArcGIS Pro 
•  "Create Pansharpened Raster Dataset"

 ENVI 
• "Pan Sharpening"
 Conclusion
• Image fusion is an important part of
Image processing
• Large amount of data limitations can be
overcome by image fusion
• Need of Image Fusion is increasing with
the progress of society
• Wide number of methods are available,
and lot of research is taking place in
image fusion
• Applications of Image fusion are limited
only by our imagination
 Reference

1. Kaur, H., Koundal, D. & Kadyan, V. Image Fusion Techniques: A Survey. Arch Computation Methods
Eng (2021)
2. Remote Sensing and GIS: Digital Image Fusion (pp.21 pages),Publisher: University Grand
Commission (UGC), MHRD, Govt of India
3. 15th European Signal Processing Conference 2007
4. www.intechopen.com/books/image-fusion-and-its-applications/image-fusion-for-remote-sensing-
applications
Thank you