This study investigated the feasibility of using near infrared spectroscopy technology to predict color and chemical composition in the heat-treated bamboo processing industry. The quantitative presentations of the changes in the chemical components were discussed using the difference spectra method of the 2nd derivative NIR spectra of the heat-treated bamboo samples. Then, the relationships between the color changes of the heat-treated bamboo and its near infrared spectra were constructed using the changes in the chemical components of the bamboo samples during the heating process. The prediction of color and chemical composition of both the outer and inner sides of the heat-treated bamboo surface were constructed using partial least squares regression method combined with a leave-one-out cross-validation process. Then, the results were validated by independent sample sets. The proposed prediction models were found to produce high r2P (above 0.93), RPD (above 3.13), and low RMSEP for both the outer and inner sides of the heat-treated bamboo samples. These studies’ results confirmed that the proposed models, especially outer side models, were perfectly suitable for the in-process inspections of the color and chemical content changes of heat-treated bamboo.
© 2020 The Author(s)PDF Article