Comparison of focal plane array FTIR pixel binning size for the nondestructive determination of cotton fiber maturity distributions

Authors: Santiago Cintron, Michael; Hinchliffe, Doug; Hron, Rebecca

Submitted to: Fibers and Polymers
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/20/2023
Publication Date: 3/9/2023
Citation: Santiago-Cintron, M., Hinchliffe, D.J., Hron, R. 2023. Comparison of focal plane array FTIR pixel binning size for the nondestructive determination of cotton fiber maturity distributions. Fibers and Polymers. 24:1473-1482. https://doi.org/10.1007/s12221-023-00149-0.
DOI: https://doi.org/10.1007/s12221-023-00149-0

Interpretive Summary: The current study further examines the use of a detector array for examining the maturity of small cotton fiber bundles. The detector array in the system was adjusted to examine 140 points in the sample, an improvement over a previous study that examined 36 points. A linear regression comparing measurements from the array system with Cottonscope maturity determination for 30 cotton samples produced an R2 value of 0.96, suggesting a strong relationship between both sets of determinations. Measurements were quick, required little sample and were not destructive to the sample. A bias examination of the 140 detection points across four calibration cottons suggests little recurring bias for most of the detection points Taken together, these results support the use of the IR detector array system for the nondestructive examination and imaging of cotton maturity distributions.

Technical Abstract: The current study further examines the use of a Focal-Plane Array (FPA) imaging system for making maturity determinations in small cotton fiber bundles. The pixel array in the system was adjusted to examine 140 points per sample, an improvement over a previous study that examined 36 points per sample. A linear regression comparing FPA and Cottonscope maturity determination for 30 cotton samples produced an R2 value of 0.96, suggesting a strong correlation between both sets of determinations. Additionally, Gaussian fits for maturity histograms included 560 sampling points per sample and mirrored maturity distributions observed with the Cottonscope. Measurements were quick, required little sample and were not destructive to the sample. A bias examination of the 140 pixels across four calibration cottons suggests little recurring bias for most of the pixels, an indication that the observed sample variability is inherent to the sample. Taken together, these results support the use of the FPA IR system for the nondestructive examination and imaging of cotton maturity distributions.