Hyperspherical Direction Cosine Transformation for Separation of Spectral and Illumination Information in Digital Scanner Data

Separation of topography-induced illumination effects and spectral information (cover type) in digital scanner data can be accomplished by projecting measurement vectors onto a hypersphere. The algorithm consists of calculating the radius R of the measurement vector X and its hyperspherical direction cosines Y for each measurement bands In X vector R = 1 and Y = -. Machine classification of the data is greatly enhanced because spectral information R (cover type)homina&s the variance in the transformed data, while topography-induced illumination effects dominate original, untransformed data. This also offers significant improvements in visual analysis and possible advantages in multitemporal analysis.