Abstract A scientifically explained color normalization process for producing image and information arrays which correspond to human color perception and description and are substantially invariant under varying conditions of illumination intensity and measurements geometry. New radiation measurements obtained by a three waveband sensor (10) are normalized by a computer or equivalent device using an individual waveband normalization algorithm (12) which divides each measurement by the mean of about ten thousand surrounding measurements, thus extracting the maximum resolution relational reflectance factor. The combined three waveband normalized array is presented as a color-coded alphanumeric array in terms of MUNSELL hue, value and chroma, using a cylindrical coordinate transformation algorithm (14) and a MUNSELL conversion algorithm (16). The normalized array is also presented using a balanced image display device (18), based on either additive or subtractive primaries.
A color normalization process comprising the traditional steps of (a) defining a plurality of radiation wavebands, (b) performing spatial array measurements of an object or objects in each of these wavebands to produce arrays of data, (c) normalizing said arrays to remove unwanted variation, (d) converting the normalized data to a standard color system, and (e) presenting the results as optical images and corresponding information arrays; the improvement comprising two elements, (1) a means in step (c) for optimizing the extraction of relational information through division of each measurement by the mean of about 10,000 surrounding measurements; and (2) a means in step (d) for converting said relational information to the MUNSELL color system.
Claim 2
Dependent Claim
114 words
A color normalization process comprising the traditional steps of (a) defining a plurality of radiation wavebands, (b) performing spatial array measurements of an object or objects in each of these wavebands to produce arrays of data, (c) normalizing said arrays to remove unwanted variation, (d) converting the normalized data to a standard color system, and (e) presenting the results as optical images and corresponding information arrays; the improvement comprising two elements, (1) a means in step (c) for optimizing the extraction of relational information through division of each measurement by the mean of about 10,000 surrounding measurements; and (2) a means in step (d) for converting said relational information to the MUNSELL color system.