A physics-based remote sensing data fusion approach

An important use of remote sensing data fusion is in retrieval of surface parameters form multisensor data. This requires using collocated observations from a least two sensors at pixel level. A method of multisensor data collocation has been developed to overlay each footprint form a coarse-resolution sensor onto a coincident finer resolution imagery data. Statistics of observations from the fine resolution data can then be correlated to the corresponding radiometric observation from the coarse resolution data. The method has been used to evaluate results from physical and empirical models to retrieve surface parameters (such as sea ice concentration), and also to develop /or modify models to calculate such parameters form a coarse resolution sensor, using subpixel information available from the fine resolution data, examples are presented to demonstrate these two applications and to emphasize the importance of using data fusion in physical modeling of remote sensoring data.