Assessment of recovered cotton fiber and trash contents in lint cotton waste by ultraviolet/visible/near-infrared reflectance spectroscopy

Authors: Liu, Yongliang; Gamble, Gary; Thibodeaux, Devron

Submitted to: Near Infrared Spectroscopy Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/27/2010
Publication Date: 10/1/2010
Citation: Liu, Y., Gamble, G.R., Thibodeaux, D.P. 2010. Assessment of recovered cotton fiber and trash contents in lint cotton waste by ultraviolet/visible/near-infrared reflectance spectroscopy. Near Infrared Spectroscopy Journal. (18)4:239-246.

Interpretive Summary: In the cotton industry, each processing stage produces a considerable amount of cotton waste or trash, which still contains some degree of cotton fiber with similar good quality to those in the bale. To assess the trash content in lint cotton, a number of instrumental methods have been developed. They include the Shirley Analyzer, a destructive gravimetric method, and high volume instrumentation (HVI), a rapid geometric method. In addition, researchers have been developing various techniques, non-spectral and spectral based, for use in the detection and identification of trash in lint cotton. Among them, NIR spectroscopy is an alternative technique due to the speed, low cost, ease of use, and potential on-line/off-line implementations. It has been successfully applied for the quantitation of reducing sugars on cotton surface and for the prediction of cotton color and physical attributes. In this approach, cotton wastes were scanned in the region of 220-2200 nm and the corresponding reference values were determined by Shirly Analyzer. Partial least squares (PLS) regression models were developed in various spectral ranges and then compared. The overal result indicated that NIR prediction of visible trash and cotton fiber in cotton waste is limited to screening purpose for probable reasons of heterogeneous trash distribution, relatively small sampling area, and gravimetric reference method. The outcome provides cotton fiber, cotton ginning, agricultural engineers and researchers a new sight in applying both optical visible/NIR and imaging spectroscopy for rapid and routine determination of cotton trash.

Technical Abstract: Lint cleaning at cotton processing facilities is performed in order to reduce the non-lint materials to the smallest level with nminimal fiber damgae. The resultant waste contains some degree of cotton fiber having equal quality to the fiber in the bale, and hence is of great concern for operating cost and profit. Traditional methods for measuring the non-lint material or trash in cotton industry, including Shirley Analyzer and high volume instrumentation (HVI), are labor intensive and time consuming. UV/visible/NIR spectroscopy, a rapid and easy sampling technique, was examined for its feasibility in determining the relative proportions fo cotton fiber and trash in lint cotton waste. Cotton waste was scanned in the region of 220-2500 nm and the reference value was measured by Shirly Analyzer. Partial least squares (PLS) regression models were developed in various spectral ranges and then compared. Though there were obvious spectral differences in visible and NIR regions between trash and cotton fiber, the model perofrmance from a narrow NIR region of 900-1700 nm was nearly equivalent to that from the full 226-2496 nm spectral region. Meanwhile, simple 2-band difference algorithms utilizing 2 unique bands were develped and alsoi suggested the effectiveness of two NIR bands at 900 and 1135 nm in the assessment of fiber and trash components in cotton waste. The overall reslut indicated that NIR prediction of visble trash and cotton fiber in cotton waste is limited to screening purpose for probable reasons of heterogeneous trash distribution, relatively small sampling area, and graimetric reference method.