Spatial resolution enhancement of SSM/I data: vegetation studies of the Amazon Basin

The relatively low resolution of the Special Sensor Microwave/Imager (SSM/I) radiometer limits its utility in vegetation studies. However, resolution enhancement techniques can be used to ameliorate this limitation. To this end the Backus Gilbert inversion (BGI) technique and a modified form of the Scatterometer Image Reconstruction (SIR) algorithm are investigated as methods to create enhanced spatial resolution images from SSM/I data. A new method for generating cloud-free composite images is presented. The utility of the composite images is illustrated through a tropical vegetation discrimination study of general vegetation classes. An overall discrimination accuracy of 81.4% is achieved for 14 classes with most of the misclassification within broader vegetation categories.