Prediction of optimal cooking time for boiled potatoes by hyperspectral imaging

Abstract

Cooking of potatoes causes changes in the microstructure and composition of starch. These changes affect the interaction of light with the starch granules at different regions inside the potato. In this research, the potential of hyperspectral imaging in the wavelength range 400-1000 nm in combination with chemometric tools and image processing for contactless detection of the cooking front in potatoes has been investigated. Partial least squares discriminant analysis (PLSDA) was employed to discriminate between the pixel spectra for the cooked regions and those for the remaining raw regions. In a next step image processing techniques were applied to detect the cooking front in the images obtained by the PLSDA pixel classification. From each of the resulting images with detected cooking fronts, the ratio of the remaining raw part area over the total potato area was then calculated. Finally, the effect of the cooking time on this ratio was modeled to be able to predict the optimal cooking time. The results illustrate the potential of hyperspectral imaging as a process monitoring tool for the potato processing industry.