Routine NIRS analysis methodology to predict quality and safety indexes in spinach plants during their growing season in the field.

Abstract

Cultural practices and harvesting in spinach plants should be based not only on subjective indexes such as freshness and green colour, which are both related with the visual appearance of the plants, but also on objective indexes that can be quantified non-destructively. The aim of this research was to develop a methodology based on the use of near infrared spectroscopy to monitor routinely the growth process of the spinach plants in the field. Using the MicroNIR™ OnSite-W spectrophotometer, which is ideally suited for in situ analysis, 261 spinach plants were analysed. Initially, calibration models for dry matter, soluble solid and nitrate contents were developed using 1 spectrum per plant for dry matter content, and nine spectra per plant for the other two parameters. These models were then validated using the same number of spectra per plant as for calibration purposes. After that, to establish a procedure more suitable to routine analysis in the field, the models were validated with only one spectrum per plant and the suitability of the predictions was measured considering the global and neighbourhood Mahalanobis distances, whose control limit values were defined as inferior to 4.0 and 1.0, respectively. The results showed that once the calibration models were developed, only one spectrum per plant was enough to predict dry matter and nitrate contents successfully. Therefore, the methodology developed will allow us to monitor in real time the complete growth process and to take decisions about spinach cultivation based on internal quality and safety indexes.

Plant phenotyping relevance

ホウレンソウ個体の内部品質・安全性指標をNIRSで非破壊推定する方法を開発し、スペクトル数を変えた検証も行っており、植物形質の取得法が中心である。

Code and data availability

The supplied blocks describe an NIRS study of spinach plants (MicroNIR OnSite-W spectra, DMC/SSC/nitrate calibrations) but contain no public phenotype dataset, spectral data deposit, author code, or trained model release. All URLs in the text are bibliographic references (cited prior work, manuals, a UC Davis factsheet

No evidence-backed public reproduction asset is currently recorded.