A Study of SAR Despeckling Methods

Authors

  • Muhammad Haris COMSATS Institute of Science & Technology, Islamabad, Pakistan
  • Mahmood Ashraf Federal Urdu University of Arts, Sciences & Technology Islamabad, Pakistan
  • Faraz Ahsan University of Islamabad, Institute of Engineering and Sciences
  • Ahmed Athar COMSATS Institute of Information Technology, Islamabad
  • Memoona Malik COMSATS Institute of Information Technology, Islamabad

Keywords:

Synthetic Aperture Radar (SAR), Optical Sensors, Radars, Speckle, Noise, Local Filters, Non-Local Filters, Wavelets, Transforms

Abstract

To capture geographic images, optical sensors were used in past to analyze and extract information for numerous purposes ranging from mineral resources to survelance. But the limited availability of optical sensors uses led to radar-based sensors for continuous data capturing. In which, neither time constraint was involved nor the hindrance by atmospheric factors. The image captured with the synthetic aperture radars undergoes the phenomena of speckle which is a granular noise triggered by the constructive or destructive superstition of received signals. The result of this interference affects the image quality and the retrieval of the information. To cope with this issue, numerous strategies were defined from local filters to non-local filters, transforms and combined solution through transforms and non-local filters as hybrid techniques. The studies of local, non-local and transforms were efficient but each having their limitation posed the question for improvement in denoising the images. The experiments conducted showed that hybrid solutions outperform their predecessors for which this study is conducted to analyze and evaluate the techniques and explore the factors in hybrid techniques that can efficiently despeckle the synthetic aperture radar images by maintaining the image quality and details.

Downloads

Published

2019-01-01