Skip to main content
ARS Home » Southeast Area » New Orleans, Louisiana » Southern Regional Research Center » Cotton Structure and Quality Research » Research » Publications at this Location » Publication #218667
Title: INSTRUMENTAL and OPERATIONAL IMPACTS on SPECTROPHOTOMETER COLOR MEASUREMENTS

Author
item Rodgers Iii, James
item Thibodeaux, Devron
item Cui, Xiaoliang
item MARTIN, VIKKI - COTTON INCORPORATED
item WATSON, MICHAEL - COTTON INCORPORATED
item KNOWLTON, JAMES - USDA AMS COTTON PROGRAM

Submitted to: Journal of Cotton Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/17/2008
Publication Date: 12/1/2008
Citation: Rodgers III, J.E., Thibodeaux, D.P., Cui, X., Martin, V., Watson, M., Knowlton, J. 2008. INSTRUMENTAL and OPERATIONAL IMPACTS on SPECTROPHOTOMETER COLOR MEASUREMENTS. Journal of Cotton Science. 12:287-297.

Interpretive Summary: Color measurements for the classing of U.S. cottons have been performed on the Uster® High Volume Instrumentation (HVI) instrument for several years. Two color parameters specific to cotton—Rd (reflectance) and +b (yellowness)—are used to express the color of cotton. Since Rd and +b do not readily relate to other well known and globally recognized color systems (e.g., L*a*b* or CIELAB), a program with the Agricultural Marketing Service (AMS) was implemented to evaluate and validate the relationships of Rd and +b to a globally recognized color system (L*a*b*) and to investigate the impacts of key instrumental and operational variables on the color results. The samples analyzed in this phase of the program were color tiles, AMS standard tiles, and AMS standard cotton fiber batts and were measured on 7 color spectrophotometers (bench-top and portable). Strong L*<->Rd and b*<->+b correlations were verified and validated on all color units. Thus, the use of L*a*b* for relating globally recognized color parameters from a color spectrophotometer to the HVI’s Rd and +b color parameters was validated. The primary variable that impacted the color agreement between units was the use of HVI glass in front of the sample. The use of glass often resulted in large DE* differences in color results between the different color instruments, especially for the portable color units. L* was the color parameter that was most impacted by the use of glass in the tile and cotton fiber color measurements. The use of glass leads to a “shift” in the sample’s reflectance spectrum (%R), and these shifts in %R, and the use of only 2 color filters in the HVI unit, lead to the observed differences in L*<->Rd and b*<->+b linearity and correlations when glass is used in the measurement.

Technical Abstract: Color measurements for the classing of U.S. cottons have been performed on the Uster® High Volume Instrumentation (HVI) instrument for several years. Two color parameters specific to cotton—Rd (reflectance) and +b (yellowness)—are used to express the color of cotton. Since Rd and +b do not readily relate to other well known and globally recognized color systems (e.g., L*a*b* or CIELAB), a program with Agricultural Marketing Service (AMS) was implemented to evaluate and validate the relationships of Rd and +b to a globally recognized color system (L*a*b*) and to investigate the impacts of key instrumental and operational variables on the color results. The samples analyzed in this phase of the program were color tiles, AMS standard tiles, and AMS standard cotton fiber batts and were measured on 7 color spectrophotometers (bench-top and portable). Strong L*<->Rd and b*<->+b correlations were verified and validated on all color units. Very good color unit agreement between the bench-top units and moderate agreement for the portable units were obtained for L*, b*, and DE* when external glass is not placed in front of the tile sample at the measurement port. Very similar linear relationships were obtained for L*<->'Rd and b*<->+b, regardless of the bench-top unit used. Thus, the use of L*a*b* for relating globally recognized color parameters from a color spectrophotometer to the HVI’s Rd and +b color parameters was validated. The primary variable that impacted the color agreement between units was the use of HVI glass in front of the sample. The use of glass often resulted in large DE* differences in color results between the different color instruments, especially for the portable color units. L* was the color parameter that was most impacted by the use of glass in the tile and cotton fiber color measurements. Distinct differences were observed between the different spectrophotometers for L* and L*'Rd correlation agreements when glass is used, but the L* differences and L*'Rd correlation agreements were usually small when no glass is used. The differences in b*<->+b were often similar overall for bench-top color units whether or not the glass was used in the color measurement. The use of glass leads to a “shift” in the sample’s reflectance spectrum (%R), and the magnitude of the shift was not linear over the entire spectrum for a given spectrophotometer and was not the same for different color spectrophotometers. These shifts in %R, and the use of only 2 color filters in the HVI unit, lead to the observed differences in L*<->Rd and b*<->+b linearity and correlations when glass is used in the measurement.