Page Banner

United States Department of Agriculture

Agricultural Research Service

Research Project: DEVELOP, ENHANCE AND TRANSFER GIN TECHNOLOGY TO IMPROVE FIBER QUALITY AND PROFITS Title: Cotton Fiber Quality Characterization with Vis-NIR Reflectance Spectroscopy: Toward an Optimal Sensor

Authors
item Ge, Yufeng -
item Thomasson, J. Alex -
item Sui, Ruixiu

Submitted to: National Cotton Council Beltwide Cotton Conference
Publication Type: Proceedings
Publication Acceptance Date: January 15, 2010
Publication Date: April 30, 2010
Repository URL: http:////www.cotton.org/beltwide/proceedings/2005-2010/data/conferences/2010/papers/10806.pdf#page=1
Citation: Ge, Y., Thomasson, J., Sui, R. 2010. Cotton Fiber Quality Characterization with Vis-NIR Reflectance Spectroscopy: Toward an Optimal Sensor. National Cotton Council Beltwide Cotton Conference. CD ROM p. 605-610.

Interpretive Summary: Cotton fiber quality is an important issue in cotton industry. Breeders develop genetic lines that give rise to superior lint quality. Producers and ginners preserve high-quality fibers through advanced harvesting and ginning technologies so that the economic value of cotton can be maximized. Specific cotton quality is also anticipated by textile processors as it will improve the efficiency of spinning and quality of dyeing. Two systems currently used for cotton fiber quality measurement and classification are High Volume Instrument (HVI) and Advanced Fiber Information System (AFIS). HVI reports several bulk fiber quality parameters including color and leaf grades, micronaire, fiber length, strength, uniformity, etc. AFIS requires a lesser volume of fibers to be tested, and it measures additional fiber quality parameters (e.g., short fiber content and nep counts). However, both HVI and AFIS are not able to make a real-time measurement. In cotton production and processing areas there are needs to determine cotton fiber quality in real time. The long term goal of this study is to develop optical sensors to measure cotton fiber quality in real time. The specific objectives of the study reported in this article were to: (1) assess the performance of the Vis-NIR (visible-near infrared) method for predicting cotton fiber quality with different calibration methods, and (2) determine useful spectral bands and bandwidths for predicting various fiber quality parameters. Sixty seed cotton samples of two varieties were handpicked and ginned with a laboratory-scale saw-type gin. Ginned lint samples were measured with a spectrophotometer in a wavelength range from 400 to 2500 nm. A portion of the lint samples was subjected to HVI measurement of six fiber quality parameters: micronaire, length, strength, uniformity, brightness (Rd), and yellowness (+b). Two methods, band-averaging (BA) and discrete wavelet transform (DWT), were used to preprocess the spectral data. Calibration models for each fiber quality parameter were developed with multiple linear regression and partial least squares regression. Spectral wavebands which were involved in the models and their bandwidths were identified. Being developed with DWT method the calibration model for micronarie involved four spectral wavebands (590, 704, 1632, 1976 nm), and the model for +b had two wavebands (590, 1496 nm). With the BA method there were six bands (430, 1470, 1590, 1650, 2070, 2230 nm) in the model for micronarie and four bands (410, 790, 810, 1750 nm) for +b. Among all six fiber quality parameters, micronaire and +b could be most successfully predicted, with R2 greater than 0.80 and 0.70, respectively. Results of this study indicated that development of optical sensors based on Vis-NIR reflectance of cotton lint appears promising.

Technical Abstract: The objectives of this research are to (1) assess the performance of the Vis-NIR method for predicting cotton fiber quality parameters with different calibration methods, and (2) determine useful spectral wavebands and bandwidths for predicting various fiber quality parameters. This study is directed toward the development of optoelectronic sensors to measure cotton fiber quality in real time in situ. Sixty seed cotton samples of two varieties were handpicked and ginned with a laboratory-scale saw-type gin. Ginned lint samples were measured with a Cary 500 UV-Vis-NIR spectrophotometer in a wavelength range from 400 to 2500 nm. A portion of the lint samples was subjected to High Volume Instrument (HVI) measurement of six fiber quality parameters: micronaire, length, strength, uniformity, brightness (Rd), and yellowness (+b). Two methods, band-averaging (BA) and discrete wavelet transform (DWT), were used to preprocess the spectral data. Calibration models for each fiber quality parameter were developed with multiple linear regression (MLR) and partial least squares regression (PLSR); and the performance of the models was assessed with cross validated root mean squared error (RMSEcv) and coefficient of determination (R2) between predicted and measured values. Among all six fiber quality parameters, micronaire and +b can be most successfully predicted, with R2 greater than 0.80 and 0.70, respectively. Prediction of length, uniformity, and strength was moderately successful, with R2 ranging from 0.40 to 0.70. Prediction of Rd was poor (R2 < 0.4). More interestingly, the RMSEcv of the calibration model for most fiber quality properties approaches the measurement accuracy of HVI instruments specified by USDA-AMS, indicating that a Vis-NIR optoelectronic sensor could perform with an acceptable level of measurement accuracy. Among the different model calibration methods, DWT-MLR generally performed better than BA-MLR and BA-PLSR, which can be attributed to the fact that DWT considers wavebands of varying bandwidths for incorporation into the model. Results of this study indicated that development of optoelectronic sensors based on Vis-NIR reflectance of cotton lint appears promising.

Last Modified: 10/21/2014
Footer Content Back to Top of Page