Chromaticity changes from isoluminous techniques used to enhance multispectral remote sensing data

Abstract

Isoluminous (equal-brightness) transformations are useful for enhancing small color differences in displays of multispectral photographic data, and for removing unwanted total radiance variations in multispectral scanner data. Several three-band isoluminous transformations were investigated mathematically to determine if they cause confusing color changes. The photographic technique of Yost and Wenderoth (1967, 1970, 1974), nonlinear in single-band functions, is shown to be strictly isoluminous only for grey shades; departure from isoluminosity is negligible for nonsaturated colors; the transformation causes an increase in chromatic saturation. A new transformation, linear in single-band functions, which is strictly isoluminous, can cause chromatic confusion unless original brightness variations are small. Another new transformation, linear in three-band functions, is also strictly isoluminous, but causes chromatic confusion. Its non-linear counterpart, ratio normalization, is found to be theoretically optimal because it is strictly isoluminous and chromaticity invariant. Photo-masks for ratio normalization in additive color projection could be produced by sequential exposure to all band records.