Optical color analysis process

• US 4,812,904 A
• Filed: 08/11/1986
• Issued: 03/14/1989
• Est. Priority Date: 08/11/1986
• Status: Expired due to Fees

• Current Assignee: Megatronics, Inc.
• Original Assignee: Megatronics, Inc.
• Inventors: Shoemaker, Martin L., Maring, Richard B.
• Primary Examiner(s): Groody, James J.
• Assistant Examiner(s): Peng, John K.
• Application Number:
  US06/895,499
• Time in Patent Office: 946 Days
• Field of Search: 358/75, 358/80, 358/76, 358/107, 382/17, 382/52, 364/526
• US Class Current: 348/135
• CPC Class Codes: G01J 2003/466 Coded colour; Recognition of predetermined colour; Determining proximity to predetermined colour
• CPC Class Codes: G01J 3/10 Arrangements of light sources specially adapted for spectrometry or colorimetry
• CPC Class Codes: G01J 3/50 using electric radiation detectors

Abstract

A process for color analysis and comparison among reference and test samples for use in quality control applications or the like comprises illuminating each sample under predetermined lighting conditions, scanning the sample with a color video camera, digitalizing the video signal output to produce a digitized signal representative of the components of the color values, preferably the red-green-blue and brightness values ("RGBW"), for each pixel representative of the viewed signal, and reporting and storing the digitized data for subsequent analysis, comparison, display and printout. For comparison purposes, the pixel color values for the samples are analyzed and compared statistically to determine if the reference and test sample match. In one test, the mean of the pixel color value for each sample is ascertained and the test sample is analyzed to determine if its mean is within a tolerance limit for the reference sample expressed in terms of standard deviation values. Various statistical tests provide useful information.

Claims

1. A process for analyzing colors in a viewing region in an object comprising illuminating the viewing region in predetermined light, scanning the viewing region with a video camera so as to produce an electronic video output for the color values at each pixel location representative of the viewing region, converting the video output into digital data for individual components of the colors, reporting and storing the digitized values in a computer memory, analyzing the digital component values to determine the means value of the component values and the extent of variation of component values from the mean value;

and selecting a predetermined variation from the mean value as a tolerance limit for comparing other colors, the tolerance limit being determined as a function of the deviation of the component values from the mean.

2. A process for analyzing colors in a viewing region in an object comprising illuminating the viewing region in predetermined light, scanning the viewing region with a video camera so as to produce a electronic vide output for the color values at each pixel location representative of the viewing region, converting the video output into digital data for individual red, green and blue ("RGB") components of the colors, reporting and storing the digitized values in a computer memory for subsequent analysis of the pattern of variation of the color values among the pixel locations, for comparison with other colors and for display, the object being viewed on a background in the viewing area and the background alone being scanned with the video camera before viewing the object, the color of the video output for the background being converted to RGB digitized values as background values and stored in the same manner as the output for the object, the background values being compared with the viewing area for the object and cancelled from the digital data for the viewing area containing the object, the resulting values reported being limited to the object and having eliminated therefrom any background values.

3. A process according to claim 2 and further comprising modifying color values to adjust for any changes in the color detection consistency of the video camera over a period of time by determining the difference between measured color values for an object of know color and previously determined digitized color values for the known color and modifying all subsequently measured color values by the amount of the difference.

4. A process for comparing colors of different objects comprising analyzing colors in a viewing region in one object by illuminating the viewing region in predetermined light, scanning the viewing region with a video camera so as to produce an electronic video output for the color values at each pixel location representative of the viewing region, converting the video output into digital data for individual components of the colors, reporting and storing the digitized values in a computer memory;

repeating the foregoing process for another object; and

comparing the digitized color component values of one object with the other object, the color component values for viewing regions on the two objects having been detected and stored separately, the difference between the mean color component values being determined so as to ascertain the degree of difference between the two colors for each of the color components, the difference indicating the amount of change necessary for each component in order to make the overall colors of the objects the same.

5. A process according to claim 4 wherein the color components for the viewing region for each object are statistically analyzed, and a tolerance based on standard deviation from a mean value is determined for color components in one object, the color component values of the other object are compared with the tolerance values of the one object to determine if the other object is within the tolerance limits, a match being indicated where the color component values fall within the tolerance limits.

6. A process according to claim 4 wherein tolerances based on standard deviation ranges are determined for each color component, the color components including the colors red, green, blue and white ("RGBW"), the other object being acceptable only if all of the component colors of the other object are within the tolerances of the component colors for the one object.

7. A process according to claim 4 wherein the stored digital can be displayed and printed out.

8. A process according to claim 4 wherein the separate color values stored include any of the color values for various viewing regions on the same or different objects, color values for reference and test objects, and color values for background, any of the color values being available for reference.

9. A process according to claim 4 wherein the color values for the objects represent reference and test objects, the color value comparison being displayed on a color monitor, with the display showing the actual color of the test object where the test color differs from reference object color by more than a preselected tolerance level, the displayed color of the test object being a different color from the actual color where the color of the test object matches the color of the reference object, the different color being selected so as to represent a color match between two compared objects.

10. A process for analyzing the color of a viewing region in a reference object comprising sampling individual color values at each of a large number of discrete locations on the viewing region, separating the color values into one or more component values and converting the values into numeric values, statistically comparing the numeric values of the components for the discreet locations on the viewing region so as to provide a statistical analysis of the variation in the color value in the viewing region, the analysis providing information about the nature and extent of color variation over the viewing region and not just a value reflecting an average color, the color values being sampled with a video camera, the discrete locations being the areas covered by individual pixels of the camera, the video signal being converted into a digitized electronic signal representing the numeric value of each color component for the pixel locations, the digitized values being compared by statistical analysis to determine the extent of variation of color values in the viewing area, the digitized color values being stored electronically for subsequent reference, the digitized color values being analyzed in accordance with statistical principles of standard deviations.

11. A process according to claim 10 wherein the digitized color values are analyzed to determine the mean value and for deviation from the mean in terms of standard deviations, With a tolerable deviation being determined in terms of standard deviation values.

12. A process for comparing colors in different viewing regions comprising analyzing the color of a viewing region in a reference object comprising sampling individual color values at each of a large number of discrete locations on the viewing region, separating the color values into one or more component values and converting the values into numeric values, statistically comparing the numeric values of the components for the discreet locations on the viewing region so as to provide a statistical analysis of the variation in the color value in the viewing region, the analysis providing information about the nature and extent of color variation over the view region and not just a value reflecting an average color, the color values being sampled with a video camera, the discrete locations being the areas covered by individual pixels of the camera, the video signal being converted into a digitized electronic signal representing the numeric value of each color component for the pixel locations, the digitized values being compared by statistical analysis to determine the extent of variation of color values in the viewing area, the digitized color values being stored electronically for subsequent reference, the color in the viewing region for the reference object being compared with a color value for a second viewing region for a second object, with the color in the second viewing region being sampled and digitized in the same manner as the color for the reference object, the second viewing region color being compared with the reference object viewing region color by a statistical comparison of the respective color values.

13. A process according to claim 12 wherein the video camera is calibrated with a reference object of known color, digital values of the known color being stored in a computer memory for reference, the extent of variation in color responsiveness of the camera being determined by a comparison between the known color values and the color values determined by actual measurement, the signal values being adjusted to compensate for any detected variation in the color responsiveness of the camera from the known color values. terms of standard deviation values.

14. A process according to claim 12 wherein the color values are statistically compared by means of one of the statistical tests consisting of the t-test, F-test, and the Kolmogorov-Smirnov test.

15. A process according to claim 12 wherein the process is performed to determine if the second color is acceptably close to the reference object color, a standard deviation being selected for the reference object color that is representative of the extent of color difference that is acceptable, the second color being analyzed to determine if the mean of the second color is within the limits of the acceptable standard deviation.

16. A process according to claim 15 wherein the color values are further statistically compared to determine if the mean value of the color of the reference object comes within a preselected standard deviation from the mean value of the second object color.

17. A process according to claim 12 wherein the video signal is converted to digitized values of the individual red, green, blue ("RGB") component values of the object color, and the RGB digitized values are the component values that are stored and analyzed.

18. A process according to claim 17 wherein the color values include a whiteness signal representative of the brightness of the colors in the signal, such that the signal becomes an RGBW signal, the value of the brightness component constituting one of the values reported and stored as being indicative of the color of the viewing region of the object.

19. A process according to claim 18 wherein the whiteness value of the video signal reflects the sum of the RGB values of the object.

20. A process according to claim 18 wherein the RGBW color component values are reported on an RGB color monitor to show the actual video appearance of each of the color components of the object, such a display permitting visual color comparison so as to confirm the accuracy of the electronic analysis process.