Hyperspectral imaging is a powerful technique that can rapidly, accurately, and non-destructively determine the quality of agricultural products. In this study, a hyperspectral imaging system has been developed to evaluate and visualize the Brix values and moisture contents in sugarcane stalks to be used as a tool for breeding programmes. After extracting the spectral data via ENVI coding, data in the wavelength range of 450–950 nm were used to generate prediction models for Brix and moisture content via partial least squares regression. The coefficients of determination of the predictive models for Brix and moisture content were found to be 0.70 and 0.68, respectively. The root mean square errors of cross-validation were 1.28° for Brix and 1.49% for moisture content, and the performance to deviation ratios were 1.71 and 1.61, respectively. The models were applied to each pixel of the hypercube data in order to determine the distributions of Brix and moisture content within the sugarcane stalks. Both distribution mappings indicated that the Brix and the moisture content level were lower in the internode regions. The results demonstrated the feasibility of using hyperspectral imaging to visualize Brix and moisture content in sugarcane stalks. The developed method has potential applications in farming management and also in breeding programs.
Nondestructive and rapid determination of lignocellulose components of biofuel pellet using online hyperspectral imaging system
