Variations of spectral signature profiles of wet and dry targets for supporting the detection of water-leakages using satellite data

Abstract

Satellite data can be used as a valuable tool for the detection of water pipeline leakages in semiarid areas. However the use of multi-temporal satellite images for this purpose can be problematic since reflectance values may change due to phenological changes of plants, radiometric errors during the pre-processing of satellite data, etc. It is therefore important to establish a spectral signature library with "ground truth data" for different scenarios of water leakages in a control site minimizing other potential errors. For this purpose, the GER 1500 spectroradiometer was used for measuring the reflectance values of three different targets: soil, vegetation and asphalt. The targets were saturated with a specific amount of water and then several spectroradiometric measurements were taken. The narrowband reflectance values were then re-scaled to spectral bands of Landsat 5 TM and spectral signature variations were highlighted for all targets before and after moisture level were increased. Using these data, threshold values were defined in order to be used for multispectral satellite data analysis. Specifically, this data was used for detection of water leakages in pipelines in Cyprus using Landsat 5 TM images.