IEEE Proceedings of 4
th
International Conference on Intelligent Human Computer Interaction, Kharagpur, India, December 27-29, 2012
978-1-4673-4369-5/12/$31.00 ©2012 IEEE
An efficient algorithm for automatic fusion of RISAT-1 SAR data
and Resourcesat-2 Optical images
Indranil Misra, Rajdeep Kaur Gambhir, S. Manthira Moorthi, Debajyoti Dhar and R.Ramakrishnan
Data Product Software Group
Signal and Image Processing Area
Space Applications Centre (ISRO), Ahmedabad-380015
E-Mail:{indranil,rajdeep,smmoorthi,deb,rama}@sac.isro.gov.in
Abstract-
Satellite Image fusion generates single hybrid image from
a collection of input satellite images and helps us to
extract maximum information from the remotely sensed
datasets to achieve optimal spatial and spectral resolution.
The critical steps of image fusion framework are co-
registration of Synthetic Aperture Radar(SAR) data with
corresponding optical scene, enhance the images for
visual clarity and then merge the multi sensor data with a
standard fusion technique. The image fusion system
should perform all these steps in an automatic manner for
providing ease to the user. The primary attention of this
work is to examine the improvement that can be obtained
by fusion of low resolution multi spectral data obtained
from optical Resourcesat-2 platform( LISS-4MX/LISS-
III/AWIFS Sensor having 5m/24m/56m spatial
resolution) with high resolution RISAT-1 (Fine
Resolution STRIPMAP (FRS-1)/Medium Resolution
SCANSAR(MRS) mode data having 3m/18m spatial
resolution) using SAR-Optical image fusion system
discussed above. This integration of optical and SAR
images from Indian Remote Sensing satellites facilitates
better visual and automatic image interpretation. The
Maximum Likelihood algorithm is used for classification
of fused image and Resourcesat-2 multispectral data. The
quality improvement of the fused product can be observed
by comparing the classification accuracies of merged data
with original multispectral data of the same region.
Keywords: Image Fusion, Resourcesat-2, RISAT-1,
SAR, Maximum Likelihood, Image Registration
I. INTRODUCTION
Remotely sensed satellite images are of great interest in
today's world for earth observation that includes resource
assessment, vegetation profile mapping and environment
monitoring. This is made possible by large amount of
datasets acquired by different types of optical and radar
sensors data. However for many space applications the
information provided by an individual sensor can be
incomplete or less in content. The reason behind the same
is that a higher spectral resolution is often achieved at the
cost of a lower spatial resolution for a given signal to
noise ratio. In current high performance computing
system, lower resolution multispectral channels are often
complemented by a single higher resolution band, which
provides a "representation" of the desired spectral and
spatial resolution shown in Fig.1 [1]. ISRO’s remote
sensing platform constellation offers considerable scope
for selection of data with multiple spatial/spectral
resolutions.
RESOURCESAT-2 was launched in April, 2011 as a
follow up mission of RESOURCESAT-1 platform
providing LISS-3/LISS-4MX/AWIFS data [2].
RESOURCESAT-2 is now available to many user
segments for operational remote sensing based natural
resource management projects. RISAT-1 is the latest
space born microwave sensing platform in the world and
first Indian microwave imaging platform for civilian
purpose [3]. This satellite was launched in April, 2012 ,
provides data in many different modes and the data
validation exercises are being carried out by ISRO teams
before it is made available to user segments in the
country.
The image fusion are therefore useful for integrating a
high spectral resolution image with high spatial resolution
image such as RISAT-1 FRS-1/MRS mode(3m/18m) with
Resourcesat-2 LISS-4MX/LISS-3/AWIFS(5m/24m/56m).
The aim of the fusion procedure is to produce high quality
and accurate data which contains the characteristic of both
the multispectral information (object identification) and
the spatial detail (object localization and texture) [4].
II. STEPS IN AUTOMATIC IMAGE FUSION
Image fusion in automatic mode requires sequential
execution of different phases for generating quality data
products. Digital Image Processing techniques are used to
generate fused image in different stages. The main
steps/phases involved in image fusion are:
1. Speckle Suppression: The speckle noise in SAR
images is a major obstacle to interpret SAR data for
various remote sensing applications. Speckle noise arises