Cotton-color instrumentation accuracy: Temperature and calibration procedure effects

Government cotton classing offices and gins with state-of-the-art process-c ontrol systems conduct automated cotton-color measurements. Calibration of the color instrumentation involves the use of five colored ceramic tiles as references for input to a multiple-linear-regression model. In this study, two commercial color-measurement instruments recorded color values of a se t of 18 colored ceramic tiles at various time intervals. The effects of cal ibrating with fever of the calibration tiles of using a simpler calibration algorithm and of temperature fluctuation, were determined regarding instru ment accuracy. Both methods of simplifying calibration (reducing the number of tiles and using a simple-linear model) had the same effect: better accu racy in the brightness measurement (Rd) and lower accuracy in the yellownes s measurement (+b). Whereas +b is the more critical measurement, the multip le-linear model with four or five tiles is superior overall. However the di fference is not so great as to render the simple-linear model with only two tiles inadequate in a process-control system that demands greater simplici ty. Temperature changes were found to be correlated with changes in measure d color values. The measured value of Rd trended downward with increasing t emperature, but there was no predictable trend in measured +b. Use of simil ar color instrumentation in a cotton gin would require some method of compe nsation for temperature effects.