Skip to main content
ARS Home » Southeast Area » New Orleans, Louisiana » Southern Regional Research Center » Cotton Structure and Quality Research » Research » Publications at this Location » Publication #269656

Title: Near-infrared classification of cotton lint, botanical and field trash

item Fortier, Chanel
item Rodgers Iii, James
item Foulk, Jonn
item Whitelock, Derek

Submitted to: Journal of Cotton Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/6/2011
Publication Date: 2/13/2012
Citation: Fortier, C.A., Rodgers III, J.E., Foulk, J.A., Whitelock, D.P. 2012. Near-infrared classification of cotton lint, botanical and field trash. Journal of Cotton Science. 16:72-79.

Interpretive Summary: Cotton trash or foreign matter comingled with cotton lint can reduce the quality and profit margin of cotton fiber. Conventional fiber quality assessment tools such as the Uster High Volume Instrument (HVITM) and the Shirley Analyzer yields non-specific information on the identity of trash found present with the lint during production, harvesting, and processing of cotton lint. A program was implemented using a near-infrared (NIR) spectroscopy spectral library to classify botanical trash (i.e. hull, leaf, seed coat, seed meat and stem) as well as to identify field trash types, which were largely composed of polypropylene and polyethylene. The spectral library had a high degree of accuracy (98%) when pure forms of botanical and field trash were studied by applying a “top-down” method to the formation of sub-libraries.

Technical Abstract: It is advantageous to produce cotton of the highest quality while minimizing the effect and presence of trash within the cotton lint. In addition, cotton trash comingled with the lint adversely affects the quality and profit margin associated with producing, harvesting, and processing cotton. The popularity of High Volume Instrumentation (HVITM) has spread because it provides instrument-based cotton quality measurements (e.g. length, strength, length uniformity, micronaire, color, and trash content). However, this instrument does not specifically determine the types of trash present within the lint. Software was utilized to determine the specific identity of various pure samples of botanical and field trash types using Near-Infrared (NIR) spectroscopy. Results of this study reveal that NIR spectroscopy identified 100% homogeneous individual samples of botanical and field trash yielding an overall accuracy of 98%.