Ship detection products based on complex SAR data in TOPS mode
Permanent link
https://hdl.handle.net/10037/11096Date
2015-06-01Type
Master thesisMastergradsoppgave
Author
Myrnes, MarianneAbstract
With the increasing need for monitoring ship traffic in both coastal and open oceans, ship detection is an important application of synthetic aperture radar (SAR). This thesis investigates image products suitable for ship detection for complex data acquired in Terrain Observation by Progressive Scans (TOPS) mode. The research was based on studying how contrast enhancement algorithms suitable for SLC data is adaptable to complex data acquired in TOPS mode. Secondly it investigated if by processing the TOPS raw data with a different focusing scheme had any affect on the image product. The evaluation of image products were done by well known contrast measurements such as peak to clutter ratio (PCR) and target to clutter ratio (TCR). The speckle content was measured by the coefficient of variation (CV). The improved subaperture cross-correlation (SCM) algorithm developed by Brekke et al. was extended to fit the dual polarization case of Sentinel-1 IW mode. The polarimetric extension was based on the full polarimetric SCM extension proposed by Souyris et al. Single, and dual polarization SCM algorithms were compared to the Polarimetric whitening filter and Aegir-Pol. Aegir-Pol is a polarimetric fusion between co-pol and cross pol widely used by the Norwegian Defence Research Establishment. Kongsberg Spacetec Near Real Time processor named NRTSAR was used to focus TOPS raw data. The focusing was done with various bandwidths and window weighting functions.The results showed that all contrast enhancement algorithms were adaptable to SLC data in TOPS mode. SCM processed for the VH channel and for the polarimetric extension yielded best image products in terms of TCR and PCR. In this relation a small overlap between the subapertures is needed. The SCM image products experienced a minor increase in TCR and PCR with increasing focusing bandwidth and the choice of window function had most effect on TCR and PCR considering the shortest bandwidth. NRTSAR focused TOPS SLC data with maximum overlap between azimuth bursts. Using this area as an alternative subaperture was investigated, but no observation of improved CV was observed.
Publisher
UiT The Arctic University of NorwayUiT Norges arktiske universitet
Metadata
Show full item recordCollections
Copyright 2015 The Author(s)
The following license file are associated with this item: