Hard mung bean seeds pose a problem in the sprouting process as they develop mold and infect neighboring seeds. Near-infrared hyperspectral imaging combined with partial least squares discriminant analysis was applied to develop a classifying model to separate hard mung beans from normal ones. The orientation of the measured beans was found to affect the classification result. The optimal partial least squares discriminant analysis model based on all orientations resulted in a correlation coefficient (R) of 0.919 with a root mean squared error of prediction of 0.197. The non-germinative parts were mapped and were concentrated at one end of the bean. Finally, a germinability index was proposed according to the proportion of colored areas between the germinative and non-germinative parts from the hyperspectral imaging results.
Near infrared hyperspectral imaging in the range 900–1700 nm was applied to classify Para rubber sheets produced from sulfuric acid coagulated latex and formic acid coagulated latex. In total, 200 Para rubber sheets were used for the development of the classification model using partial least squares discriminant analysis. Two types of sample presentation based on the reflectance and transflectance modes were compared to classify performance. The results indicated that both modes of measurement provided good classification with 98.33% accuracy. However, the transflectance mode yielded better performance than the reflectance mode as the former model had a higher correlation coefficient (0.88), lower standard error of prediction (0.176) and fewer preprocessing methods. The best transflectance mode was based on the absorbance pre-processed using the standard normal variate and the Savitzky-Golay second derivative whereas the best reflectance mode was developed using the absorbance pre-processed with a combination of Savitsky-Golay smoothing, the standard normal variate and the Savitzky-Golay second derivative. The fewer preprocessing methods in the transflectance mode suggested that less noise was present in the individual spectra compared to the spectra gathered in the reflectance mode. There was a clear difference between the color-mapped images of Para rubber sheets coagulated using sulfuric acid and formic acid, especially with the transflectance model. Each pixel in the images could be more accurately classified by the transflectance model.