Submitted to: National Cotton Council Beltwide Cotton Conference
Publication Type: Proceedings
Publication Acceptance Date: February 25, 2010
Publication Date: June 11, 2010
Citation: Thibodeaux, D.P., Liu, Y., Gamble, G.R., Foulk, J.A. 2010. A new approach to cotton trash measurement. National Cotton Council Beltwide Cotton Conference, January 4-7, 2010, New Orleans, Louisiana. p. 1337. Interpretive Summary: The reference method for measuring cotton trash is the Shirley Analyzer (ASTM D2812 - 07 “Standard Test Method for Non-Lint Content of Cotton”). We will review approaches to utilize either the HVI or AFIS as substitute methods for measuring trash as compared with the Shirley. Preliminary results will also be presented wherein CQRS has utilized a spectrometer operating in the UV, visible, and near infrared regions of the spectrum to predict the trash content of cotton as measured by the Shirley Analyzer (SA).
Technical Abstract: Two different studies were conducted to determine the feasibility of measuring the mass of trash in a sample of cotton lint by means of a rapid/non-destructive spectroscopic method. The first involved three cottons grown in three states subjected to two different harvesting treatments and processed into yarn with waste samples collected at several points in the process line. A total of 165 samples were analyzed using a JASCO V-670 UV/visible/NIR spectrometer. Reflectance (R) readings were acquired over the 220 - 2500 nm wavelength range at 1-nm intervals and converted into log (1/R) values. The second involved three cottons grown in Texas that were subjected to nine different ginning/cleaning treatments with a total of eight reps yielding 216 different samples. A total of 175 samples were analyzed using a FOSS XDS Rapid Content Analyzer. Reflectance (R) readings were acquired over the 400- 2500 nm wavelength range at 1-nm intervals and converted into log (1/R) values. Preliminary results indicate it is feasible to obtain reasonable (R2>0.8) models using UV/VIS/NIR instruments to predict the gravimetric trash content of cotton. Even with limited sample size using a relatively simple instrument it is possible to obtain very good results (R2~0.9) with a set of samples containing a wide ranging amounts of trash. Using a more sophisticated instrument with a larger sample holder, we can obtain good results (R2~0.85) with a set of lint samples containing a relatively narrow range of non-lint content. We intend to continue this research using the FOSS to determine the impact of increased repetitions on enhancing accuracy.