Colorimetry and the multichannel image

Abstract

The use of color theory (particularly color metrics) is shown in the display of 2-D vector fields. Extant concepts in color theory are applied to data spaces to facilitate mapping from data into color. Novel color transformations are shown; one is a direct improvement over standard Landsat imagery, one is a means of showing in a single image the results of analyzing an image of arbitrarily large dimensionality, and another is an improved method of displaying geophysical attributes that reduces the number of separate displays by a factor of three. A new model for the inversion of a dot-matrix color machine is given. A method is formulated and demonstrated to convert a numerical image to halftone printed copy and avoid several color-contaminating physical steps previously necessary. Analysis of a differentiable color image machine by its Jacobian matrix is explored and then extended to discrete (non-differentiable) machines. Finally, an inverse problem is formulated -- the extraction of quantitative information from color imager, in the case of one principal continuous variable and subsidiary variable having lesser color impact.