The fusion of polarimetric AirSAR data with digital aerial photography to improve aquatic macrophtye classification in a tropical reservoir

Previous attempts to map wetland environments in northern Australia using optical satellite sensors have highlighted the difficulties in mapping the vegetation in these systems using remotely sensed data. Recently, more promising results have been achieved through the use of polarimetric AirSAR obtained during the 1996 PACRIM campaign. The mapping has been conducted on a tropical floodplain containing a reservoir, Fogg Dam, that is characterised by a permanent water supply. A maximum likelihood supervised classification was implemented on four data sets: CIR photography; a Cloude decomposed 9 band AirSAR image; the optimum 3 band combination Cloude decomposed image; and the fused data. The macrophtye classification was improved when utilising AirSAR data, with the 3 band decomposed image yielding the most accurate results