O'Grady, Damien, Leblanc, Marc, and Gillieson, David (2013) Relationship of local incidence angle with satellite radar backscatter for different surface conditions. International Journal of Applied Earth Observation and Geoinformation, 24. pp. 42-53.

PDF (Published Version) - Published Version Restricted to Repository staff only

 

34

5

Abstract

This paper examines the relationship of C-band radar backscatter from the Advanced Synthetic Aperture Radar on board the ENVISAT satellite with the local angle of incidence, whose influence on the received signal is significant, particularly in the modes of sensor operation that use the full swath of the orbit track. Linear regression is carried out for each pixel throughout a large time series of radar data over the whole of the state of Queensland, Australia, and at Great Salt Lake, Utah, USA. In the first case, the resultant coefficients are analysed for correlation against various parameters, with regolith showing the highest correlation. Class separability analysis shows the potential to use the resultant coefficients as a supplement to absolute threshold values in order to distinguish between classes of vegetation and/or geology, where cloud cover may preclude the use of optical data. It is observed that the separability between water and land is greatly higher using the slope coefficient B than using backscatter 0 , which may be of great benefit in the remote sensing of water where cloud cover is present (from which radar is largely independent). This is especially the case when considering the observed overlapping of backscat- ter values from water, with values from aeolian sand and lacustrine and alluvial sediments, rendering the use of backscatter alone problematic. In order to test the potential use of B to map water extents, the study over the Great Salt Lake compares the classification accuracy of B against that of 0 . It is found that the 0 classification misrepresents desert, salt flat and dry lake basin areas, where the B classifica- tion differentiates these regions accurately. The resultant classification achieves a kappa statistic around 0.9, which shows very high conformance. An accurate and novel method to classify water is therefore demonstrated, which awaits the launch of anticipated improved synthetic aperture radar instruments on satellite missions in the coming few years.

Actions (Repository Staff Only)

Item Control Page