Reduction of Speckle Noise in SAR Image with the Application of DSF Filter

Abstract

SAR (Synthetic Aperture Radar) is a mechanism of radar communication crafted for the
object reconstruction with the inclusion of both two and three dimension. It is based on the
motion of radar antenna over a target region. SAR has an enormous range of practicalities
which includes earth observation applications, areas of natural resource management, etc.
The former being covering areas in global monitoring, mapping, charting, and land use
planning. The latter is its application in fields of forestry agriculture, water quality
monitoring and wildlife habitat management. Being so well rounded in its functioning, there
is a vital importance in retaining the necessitated information of the SAR imagery during its
compression. The presence of inherent speckle noise demeans the quality of SAR. Multiple
filters viz. the Lee or Frost filters, based on variation in adaptive filtering coefficient, have
been utilised to restore the reflectivity of radar with the underlying assumption of the
presence of multiplicative speckle noise in the image. However, due to the occurrence of
strong scatters or the structural features (lines or edge) in processing windows, this type of
speckle filtering is not as effective as the method proposed by this paper. This is an algorithm
for the filtering using Directional Smoothing Filter (DSF). Thus, following noise removal and
information preservation, the proposed method helps segue into high data transmission rates
and archival ratios.