Pre-processing Algorithm for Rectification of Geometric Distortions in Satellite Images

Narayan Panigrahi; B. K. Mohan; G. Athithan

doi:10.14429/dsj.61.421

Narayan Panigrahi Centre for Artificial Intelligence and Robotics, Bengaluru
B. K. Mohan Indian Institute of Technology Bombay, Mumbai
G. Athithan Centre for Artificial Intelligence and Robotics, Bengaluru

DOI: https://doi.org/10.14429/dsj.61.421

Keywords: Input-to-output image transformation, Polynomial affine transformation, Ortho-rectification, Pre-processing algorithm, Image rectification, Satellite images, Image registration

Abstract

A number of algorithms have been reported to process and remove geometric distortions in satellite images. Ortho-correction, geometric error correction, radiometric error removal, etc are a few important examples. These algorithm require supplementary meta-information of the satellite images such as ground control points and correspondence, sensor orientation details, elevation profile of the terrain, etc to establish corresponding transformations. In this paper, a pre-processing algorithm has been proposed which removes systematic distortions of a satellite image and thereby removes the blank portion of the image. It is an input-to-output mapping of image pixels, where the transformation computes the coordinate of each output pixel corresponding to the input pixel of an image. The transformation is established by the exact amount of scaling, rotation and translation needed for each pixel in the input image so that the distortion induced during the recording stage is corrected.

Defence Science Journal, 2011, 61(2), pp.174-179, DOI:http://dx.doi.org/10.14429/dsj.61.421

Author Biographies

Narayan Panigrahi, Centre for Artificial Intelligence and Robotics, Bengaluru

Mr Narayan Panigrahi has received his MSc (Computer Science) from J K institute of Applied Physics & Technology, University of Allahabad and MTech (CSDP) from Indian Institute of Technology (IIT), Kharagpur, in 1991and 1999, respectively. Currently he is pursuing his PhD from IIT Bombay, Mumbai. Presently he is working at Centre for Artificial Intelligence and Robotics (CAIR), Bengaluru, on the design and development of a geographical information system (GIS). His research areas include: GIS, image processing and design and optimisation of computational geometric algorithms. He has authored two books on geographical information science and published seven research papers in various international journals.

B. K. Mohan, Indian Institute of Technology Bombay, Mumbai

Dr B. Krishna Mohan has received his PhD in Electrical Engineering from IIT Bombay, Mumbai in 1991. His area of specialization is satellite image processing and analysis. His research interests include Object based analysis of high resolution images, hyperspectral image analysis, geographic information systems, and multimedia educational content development for digital image processing. He has over 60 papers in international and national journals and conferences.

G. Athithan, Centre for Artificial Intelligence and Robotics, Bengaluru

Dr Gopalasamy Athithan has received his PhD in Physics (Neural Networks) from the IIT Bombay, Mumbai, in 1997. Presently, he is working as Outstanding Scientist at Centre for Artificial Intelligence and Robotics, Bengaluru, on the development of information security solutions. His current interests are information security management, brain modelling through neural networks, and complex dynamical systems. He has published about twenty papers in archived journals and 20 papers in national and international conferences.