Two-Dimensional TSVD to Enhance the Spatial Resolution of Radiometer Data

A reconstruction technique, based on the 2-D truncated singular value decomposition, is first proposed to enhance the spatial resolution of radiometer earth observation measurements. The technique is very computer time effective when the kernel is a 2-D tensor product. The key issue regarding the selection of the truncation parameter is addressed by the statistically based generalized cross-validation approach. Experiments undertaken on a data set consisting of both simulated and actual 2-D special sensor microwave imager radiometer measurements show the robustness of the technique against the additive noise and its effectiveness in terms of processing time. A typical 2-D radiometer scene is processed in seconds by a standard PC processor.