Abstract

Lint cleaning at cotton processing facilities is performed in order to reduce the non-lint materials to the lowest level with minimal fibre damage. The resultant waste contains some degree of cotton fibre having equal quality to the fibre 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 the cotton industry, including the Shirley Analyzer and high volume instrumentation, are labour intensive and time consuming. Ultraviolet (UV)/visible/near infrared (NIR) spectroscopy, a rapid and easy sampling technique, was examined for its feasibility in determining the relative proportions of recovered cotton fibre and trash in lint cotton waste. Cotton waste was scanned in the region of 220–2500 nm and the reference value was measured by a Shirley Analyzer. Partial least squares regression models were developed in various spectral ranges and then compared. Although there were obvious spectral differences in the visible and NIR regions between trash and cotton fibre, the model performance from a narrow NIR region of 900–1700 nm was nearly equivalent to that from the full 226–2496 nm spectral region. Meanwhile, simple two-band difference algorithms utilising two unique bands were developed and also suggested the effectiveness of two NIR bands, at 900 nm and 1135 nm, in the assessment of fibre and trash components in cotton waste. Furthermore, considering the different sampling species between the reference and spectral measurements, a 90% confidence interval was applied to remove outlier samples from the calibration and validation sets. The recalibrated models revealed the feasibility of the NIR technique for the determination of visible trash and recovered cotton fibre content in lint cotton waste.

© 2010 IM Publications LLP