Skip to main content
ARS Home » Research » Publications at this Location » Publication #219980
Title: Principal Component Analysis (PCA) Prediction of Cotton Trash

Author
item Allen, Angela
item Foulk, Jonn
item Gamble, Gary

Submitted to: Proceedings of Pittcon Meeting
Publication Type: Abstract Only
Publication Acceptance Date: 2/21/2007
Publication Date: 3/2/2007
Citation: Allen, A.M., Foulk, J.A., Gamble, G.R. 2007. Principal Component Analysis (PCA) Prediction of Cotton Trash. Proceedings of Pittcon Meeting.

Interpretive Summary: The presence of cotton trash plays a crucial role in the quality assurance of cotton fibers in the textile industry. The application of Fourier transform infrared spectorscopy as a potential tool for the determination of the type of cotton trash in cotton fibers was investigated. Cotton trash taken from various environmental conditions (same location, different variety and same variety, different location)were separated into different types (stem, leaf, and hull), reduced to smaller particle sizes (pepper size), and extracted in order to separate constituents associated with various trash types. Using the FTIR spectral data, PCA was applied in an effort to predict the type of cotton trash through extracted constituents. Previous investigations revealed an optimum calibration using a combination of preprocessing modifications to achieve classification among the selected cotton trash.

Technical Abstract: Since cotton trash adversely affects the "quality" and market grade of cotton, the spinning efficiency in mill, and the resulting manufactured fabric passed on to consumers, ginning and textile processing are performed in a manner that removes as much of the contamination as possible. However, the identification of the trash types that persists in cotton during processing is important in addressing quality concerns and the major sources of the contamination (i.e., harvesting and ginning). The application of PCA methods for FTIR calibration was invesigated as a potential means for determining trash types. Cotton trash samples taken from various environmental conditions were separated into different types, reduced to smaller particle sizes (pepper size), and extracted (methanol) in order to separate consituents associated with the various trash types. Using the FTIR spectra, a PCA model was utilized to predict the types of cotton trash present in the various varieties and to determine the existence of relationships of trash types between varieties and growing locations.