Impervious surface area extraction using simulated EnMAP imagery

The future launch of EnMAP satellite in 2019 will enable the acquirement of hyperspectral data for obtaining more accurate relevant surface parameters on a global scale. This paper will discuss the application of EnMAP in extracting impervious surface area (ISA) from urban-rural gradient scenes. Considering the need for ISA extraction in urban studies, the choice of endmembers are made of five mixed types: vegetation, high and low albedo (impervious surface area), soil, and water. This research uses the robust non-negative matrix factorization to simultaneously generate endmembers and the corresponding abundances. It has been found that the strategy of mixed endmembers is a reasonable approach in median spatial resolution spaceborne imagery. Even with the median resolution of 30m, the EnMAP derived ISA is comparable to the result from high spatial resolution (1m) airborne data.