ND-Space: Normalized Difference Spectral Mapping, with Soil and Vegetation Examples

Identification of materials based on color is complicated by independent variations in intensity. Differences in the viewing/illumination directions, shadows, soil moisture content, etc., can obscure spectral changes and complicate modeling of those changes. A normalization procedure that removes intensity variations can greatly clarify the nature of the spectral change, allowing for more effective interpretation of color change in spectral imagery. Normalized difference measures are used here because of the simplicity of the computation, the convenient scaling [−1,1], and the ease with which the normalized difference procedure can be extended to multiple dimensions. A vector pace is described providing improved discrimination among bare soils and emergent vegetation, and a 2-D model of soil-specific change in vegetation density is presented.