Debajyoti Dhar
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Debajyoti Dhar has a total of 17 co-authors. Here are the top co-authors by number of shared publications:
- Manthira Moorthi Subbiah
India Space Research Organization
4 shared publications - Indranil Misra
Signal and Image Processing Group, Space Applications Centre, Indian Space Research Organisation (ISRO), Ahmedabad, India
3 shared publications - Mukesh Kumar Rohil
Department of Computer Science and Information Systems, Birla Institute of Technology and Science Pilani, Pilani Campus, Pilani, India
2 shared publications Ankur Garg
India Space Research Organization
1 shared publication- Ashutosh Gupta
ISRO
1 shared publication Hasit Parmar
Computer Engineering Department, L. D. College of Engineering, Ahmedabad, India
1 shared publication
Latest Publications
- 6 December 2025
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 47 (1) , 399-429
Ground image processing of high-resolution thermal EOIR data from EOS-08 microsatellite mission
- 6 December 2025
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 47 (1) , 399-429
Ground image processing of high-resolution thermal EOIR data from EOS-08 microsatellite mission
- 30 November 2023
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 44 (23) , 7365-7389
SAAC-Net: deep neural network-based model for atmospheric correction in remote sensing
- 30 November 2023
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 44 (23) , 7365-7389
SAAC-Net: deep neural network-based model for atmospheric correction in remote sensing
- 8 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
DEM Generation From Mono-Multispectral Images via Uncalibrated Epipolar Rectification and 3D Registration
Generation of Digital Elevation Models (DEM) typically requires fine image matching between set of images acquired with a wide-baseline and appropriate B/H ratio. Wide baseline stereo imaging is an expensive process since it requires two or more acquisitions at the cost of nominal mono-imaging operations. Due to the wide-baseline, angular separation in views is required, which results in increased occlusion problems leading to holes in the generated DEM. Multiview images taken with wide baselines also have lower overlap area compared to images taken from narrow-baseline systems. Moreover, while making orthorectified maps, large tilts need to be corrected using processing techniques to obtain a top-down view. To alleviate these issues, we propose a novel DEM generation pipeline that utilizes the narrow-baseline information between bands of a single multispectral image. Our method can generate sub-pixel accurate disparity maps while tackling spectral information change. Final DEM is generated from the disparity by using reference Ground Control Points (GCPs) and an iterative control point registration procedure. We demonstrate the effectiveness of our DEM generation pipeline on Resourcesat-2 LISS- IV multispectral image with a BIH ratio as low as 0.018. Our results show that sub pixel disparity of the order of 1/50th of a pixel can be computed. A visual and quantitative comparison of the generated DEM with the wide-baseline DEM from Cartosat-l shows that the elevation information is estimated with good overall agreement even for a hilly region. Our technique makes it possible to extract elevation information from mono-multispectral images with unusually low B/H ratios, potentially enabling the generation of DEMs from multispectral images that would not normall be considered viable for this purpose.
- 8 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
DEM Generation From Mono-Multispectral Images via Uncalibrated Epipolar Rectification and 3D Registration
Generation of Digital Elevation Models (DEM) typically requires fine image matching between set of images acquired with a wide-baseline and appropriate B/H ratio. Wide baseline stereo imaging is an expensive process since it requires two or more acquisitions at the cost of nominal mono-imaging operations. Due to the wide-baseline, angular separation in views is required, which results in increased occlusion problems leading to holes in the generated DEM. Multiview images taken with wide baselines also have lower overlap area compared to images taken from narrow-baseline systems. Moreover, while making orthorectified maps, large tilts need to be corrected using processing techniques to obtain a top-down view. To alleviate these issues, we propose a novel DEM generation pipeline that utilizes the narrow-baseline information between bands of a single multispectral image. Our method can generate sub-pixel accurate disparity maps while tackling spectral information change. Final DEM is generated from the disparity by using reference Ground Control Points (GCPs) and an iterative control point registration procedure. We demonstrate the effectiveness of our DEM generation pipeline on Resourcesat-2 LISS- IV multispectral image with a BIH ratio as low as 0.018. Our results show that sub pixel disparity of the order of 1/50th of a pixel can be computed. A visual and quantitative comparison of the generated DEM with the wide-baseline DEM from Cartosat-l shows that the elevation information is estimated with good overall agreement even for a hilly region. Our technique makes it possible to extract elevation information from mono-multispectral images with unusually low B/H ratios, potentially enabling the generation of DEMs from multispectral images that would not normall be considered viable for this purpose.
- 18 June 2022
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 43 (12) , 4477-4516
Feature based remote sensing image registration techniques: a comprehensive and comparative review
Earth observation using remote sensing data is a trending research field across the globe. In this perspective, image registration is a mandatory data processing step for any kind of time series data analytics. Feature-based image registration is one of the prominent categories for superimposing multi-temporal and multi-modal images over each other. The paper provides a comprehensive and comparative survey of feature detection/description techniques for remote sensing images. In addition, outlier removal algorithms are explored in detail to generate putative keypoint correspondences for accurate transformation parameter estimation. The experiments are conducted on multiple remote sensing image pairs comprising visible, Synthetic Aperture Radar (SAR) and infrared images that provide diverse characteristics of the feature target. The co-registration accuracy is quantified for all possible combinations of feature detection/description with outlier removal, and best amalgamation is visually verified to check sub-pixel spatial alignment by swiping multi-temporal or multi-modal images over each other. It has been found that SIFT performs better in optical image registration, whereas A-KAZE feature detection has an upper edge at SAR image registration task. Marginalizing Sample Consensus (MAGSAC) and Mode Guided (MG) based outlier removal techniques achieves better pruning performance for multi-temporal optical and SAR images respectively. The comparative evaluation indicates that Motion Smoothness Constraint (MSC) optimization shows optimal performance for multi-modal remote sensing images. The best possible Root Mean Square Error (RMSE) achieved for multi-temporal optical images is 0.44 pixel whereas for multi-temporal SAR images, it is 0.46 pixels. The RMSE achieved for multi-modal images is 0.48 pixel using combination of SIFT with MSC.
- 18 June 2022
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 43 (12) , 4477-4516
Feature based remote sensing image registration techniques: a comprehensive and comparative review
Earth observation using remote sensing data is a trending research field across the globe. In this perspective, image registration is a mandatory data processing step for any kind of time series data analytics. Feature-based image registration is one of the prominent categories for superimposing multi-temporal and multi-modal images over each other. The paper provides a comprehensive and comparative survey of feature detection/description techniques for remote sensing images. In addition, outlier removal algorithms are explored in detail to generate putative keypoint correspondences for accurate transformation parameter estimation. The experiments are conducted on multiple remote sensing image pairs comprising visible, Synthetic Aperture Radar (SAR) and infrared images that provide diverse characteristics of the feature target. The co-registration accuracy is quantified for all possible combinations of feature detection/description with outlier removal, and best amalgamation is visually verified to check sub-pixel spatial alignment by swiping multi-temporal or multi-modal images over each other. It has been found that SIFT performs better in optical image registration, whereas A-KAZE feature detection has an upper edge at SAR image registration task. Marginalizing Sample Consensus (MAGSAC) and Mode Guided (MG) based outlier removal techniques achieves better pruning performance for multi-temporal optical and SAR images respectively. The comparative evaluation indicates that Motion Smoothness Constraint (MSC) optimization shows optimal performance for multi-modal remote sensing images. The best possible Root Mean Square Error (RMSE) achieved for multi-temporal optical images is 0.44 pixel whereas for multi-temporal SAR images, it is 0.46 pixels. The RMSE achieved for multi-modal images is 0.48 pixel using combination of SIFT with MSC.
- 1 July 2021
- Published by Elsevier in Planetary and Space Science
- Vol. 201, 105215
Phobos image enhancement using unpaired multi-frame acquisitions from Indian Mars Color Camera
- 1 July 2021
- Published by Elsevier in Planetary and Space Science
- Vol. 201, 105215
Phobos image enhancement using unpaired multi-frame acquisitions from Indian Mars Color Camera