The impact of shadow enhancement algorithms on remotely sensed images of complex urban environments

Abstract

Large portions of shadowed areas in satellite images of urban areas can affect the accuracy of classification and thus reduce an image's effectiveness in urban remote sensing applications. This is particularly acute in cities such as Hong Kong where dense high-rise buildings cast many long shadows across a variety of different surface types. One solution to this problem is to enhance shadowed areas so their spectral range becomes closer to their corresponding non-shadowed areas. In this thesis the Spectral Shape Index was used to identify shadowed areas and two techniques, Gamma Correction and Linear Correlation Correction, were applied for the enhancement of three study sites of 2.4m spatial resolution multispectral Quickbird image. The selected study sites represent typical urban types of Hong Kong, ranging from high-rise commercial to low-rise residential areas. The performance of the shadow detection algorithm and its limitations are discussed. The histograms of the corresponding non-shadowed areas, the original and the enhanced shadow areas are used to compare the spectral range. Problems associated with shadow enhancement are discussed and the possibility of using a non-linear model for enhancement is examined. The results show that Linear Correlation Correction is more suitable when applied to complex environments and the enhanced areas as in band ratios, such as NDVI, show greater similarity after enhancement. For shadows that are still darker after enhancement, a second iteration was performed and the results were examined. It was found that when the shadows are extremely dark the spectral information will be damaged and cannot be enhanced effectively.