Abstract: Systems, computer-readable media, and systems are provided for determining whether a data point indicates a presence of a shadow-covered ground point. A data point from top of atmosphere data from an imaging study of an area potentially covered by a cloud shadow is selected. At least one spectral data measurement associated with the data point, the spectral data measurement including at least one of visible, near-infrared, and short wavelength infrared data is taken. At least one of the spectral data measurement and derived spectral index is compared with a spectral data threshold, the spectral data threshold delineating between a shadow-covered ground point and a non-shadow-covered ground point. The data point is classified as one of a shadow-covered ground point and a non-shadow covered ground point based on the comparison with the spectral data threshold.

Abstract: Methods, computer-readable media, and systems are provided for determining whether a data point indicates a presence of a sub-visible cloud using visible near-infrared data and short wavelength infrared data. A data point is selected from an imaging study of an area potentially covered by at least one of visible clouds and sub-visible clouds. A presence of a sub-visible cloud is determined. The determination is made by comparing a cirrus-band reflectance of the data point with a sub-visible cirrus-band reflectance threshold. The data point is classified as a sub-visible cloud point if the cirrus-band reflectance of the data point exceeds the sub-visible cirrus band reflectance threshold.

Abstract: Methods, a computer-readable medium, and a system are provided for determining whether a data point indicates a presence of a cloud using visible near-infrared data and short wavelength infrared data. A first comparison of a cirrus-band reflectance of the data point with a threshold cirrus-band reflectance value is made, classifying the data point as a cloud point if the cirrus-band reflectance of the data point exceeds the threshold cirrus-band reflectance value. When the comparing of the cirrus-band reflectance of the data point with the threshold cirrus-band reflectance value does not classify the data point as a cloud point, a further analysis is performed, including performing a second or more comparisons of additional cloud indicators derived from at least one of the visible, near-infrared, and short wavelength infrared data with related empirically-derived, landcover-dependent thresholds for classifying the data point as a cloud point or a non-cloud point.