Positioning in spectral measurement dominates estimation performance of internal rot in onion bulbs
| Autor: | Masaru Nishino, Yohei Deguchi, Shinichi Nakano, Hiromichi Itoh, Shin-Ichiro Kuroki |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
Chemistry
business.industry 010401 analytical chemistry Near-infrared spectroscopy 04 agricultural and veterinary sciences Horticulture Onion bulb 01 natural sciences 040501 horticulture 0104 chemical sciences Bulb Absorbance Optical axis Optics Partial least squares regression Transmittance Nir spectra 0405 other agricultural sciences business Agronomy and Crop Science Food Science |
| Zdroj: | Postharvest Biology and Technology. 128:18-23 |
| ISSN: | 0925-5214 |
| DOI: | 10.1016/j.postharvbio.2017.02.001 |
| Popis: | The feasibility of visible and near infrared (VIS/NIR) transmittance spectroscopy for nondestructive discrimination of healthy and diseased onion bulbs was examined. Rottenness occurring in inner scales that were hard to diagnose by the five human senses was targeted. First, VIS/NIR spectra were acquired for six bulb positions. Subsequently, the degree of rot was measured by visual inspection of vertically cut surface based on our proposed criteria and by weighing of sectioned rotted parts. Our evaluation criteria functioned as an alternative mode of the weighing method. The basal plate of onion bulb had remarkable scattering properties. Therefore, spectral features differed not only by the rot levels but also by positioning the bulb against the sensor setup. The partial least squares regression model developed from the longitudinal position, for which the basal plate did not overlap against the system optical axis, showed the best performance and classified rotted bulbs above level 1 with 98.4% accuracy. Absorbance below 736 nm was the informative region for bulb rottenness detection. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect