A comparison of two spectral mixture modelling approaches for impervious surface mapping in urban areas

Urban change processes that have been occurring over the last decades are affecting the human and natural environment in many ways, and have stressed the need for new, more effective urban management approaches. In this context, mapping man-made impervious surfaces has received quite some attention as impervious surfaces can be used as a general indicator to quantify urban change and its environmental impact. Despite the currently available digital imagery from high-resolution satellite sensors like Ikonos and Quickbird, or from airborne cameras, spectral unmixing approaches applied on medium-resolution data from sensors like Landsat TM/ETM+ or SPOT-HRV offers interesting perspectives to map impervious surfaces for large spatial extents. While several techniques for sub-pixel impervious surface mapping have already been examined, there is a lack of comparative analysis. Our objective was to compare two spectral mixture analysis models: linear spectral unmixing versus multi-layer perceptrons (MLP). Both models were implemented in a multi-resolution framework, where reference data for model training was obtained from a high resolution land-cover classification (derived from Ikonos imagery), while the models themselves were applied on medium resolution data (Landsat ETM+). As a secondary objective, the effect of spectral normalisation on the performance of both models was assessed. The model based on MLP performed clearly better than the linear mixture model. The average absolute error of the impervious surface proportion estimate within each medium resolution pixel was 10.4% for the MLP model versus 12.9% for the linear mixture model. Spectral normalisation proved to be very useful to improve the results obtained by the linear mixture model as the mean absolute error for impervious surfaces decreased from 14.8% to 12.9% after normalisation. Its effects on the MLP model appeared to be insignificant. The outcome of this study helps to provide guidance for the selection of an approach to estimate continuous impervious surface fractions from medium resolution data.