ASABE Technical Library - Abstract
Member and Access Notice
Onion Quality Assessment Using a Near Infrared Hyperspectral Imaging System
Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan www.asabe.org
Citation: 2012 Dallas, Texas, July 29 - August 1, 2012 121341135.(doi:10.13031/2013.42213)Authors: Haihua Wang, Changying Li
Keywords: LCTF, soluble solids content, dry matter, genetic algorithm
A liquid crystal tunable filter based near infrared hyperspectral imaging system was used to assess onion internal quality (firmness, soluble solid content, and dry matter content). Reflectance, transmittance, and interactance images were collected over the spectral region from 950 to 1750 nm. The spectra were extracted from regions of interest in the spectral images, which were used to develop calibration models using the partial least square regression. For interactance images, the regions of interest were identified at 1200 nm wavelength by a gray gradient algorithm. In order to reduce the number of spectral variables and increase the performance of the model, genetic algorithm was used to select optimal wavelengths. Results showed that the reflectance mode performed the best among the three sampling modes. The soluble solid content and dry matter content could be better predicted than the firmness. The hyperspectral imaging system could be used for nondestructive evaluation of onion internal qualities
(Download PDF) (Export to EndNotes)