Skip to main content

ORIGINAL RESEARCH article

Front. Plant Sci.
Sec. Crop and Product Physiology
Volume 15 - 2024 | doi: 10.3389/fpls.2024.1456800
This article is part of the Research Topic Olive Science View all 17 articles

Predicting oil accumulation by fruit image processing and linear models in traditional and super high-density olive cultivars

Provisionally accepted
Giuseppe Montanaro Giuseppe Montanaro 1*Antonio Carlomagno Antonio Carlomagno 1Angelo Petrozza Angelo Petrozza 2Francesco Cellini Francesco Cellini 2Ioanna Manolikaki Ioanna Manolikaki 3Georgios Koubouris Georgios Koubouris 3Vitale Nuzzo Vitale Nuzzo 1
  • 1 University of Basilicata, Potenza, Italy
  • 2 Metapontum Agrobios Research Center, Lucana Agency for Development and Innovation in Agriculture, Metaponto, Italy
  • 3 Hellenic Agricultural Organization – ELGO, Athens, Greece

The final, formatted version of the article will be published soon.

    The paper focuses on the seasonal oil accumulation in traditional and super-high density (SHD) olive plantations and its modelling employing image-based linear models. For these purposes, at 7-10-day intervals, fruit samples (cultivar Arbequina, Fasola, Frantoio, Koroneiki, Leccino, Maiatica) were pictured and images segmented to extract the Red (R), Green (G), and Blue (B) mean pixel values which were re-arranged in 35 RGB-derived colorimetric indexes (CIs). After imaging, the samples were crushed and oil concentration was determined (NIR). The analysis of the correlation between oil and CIs revealed a differential hysteretic behavior depending on the covariates (CI and cultivar). The hysteresis area (Hyst) was then quantified and used to rank the CIs under the hypothesis that CIs with the maximum or minimum Hyst had the highest correlation coefficient and were the most suitable predictors within a general linear model. The results show that the predictors selected according to Hyst-based criteria had high accuracy as determined using a Global Performance Indicator (GPI) accounting for various performance metrics (R 2 , RSME, MAE). The use of a general linear model here presented is a new computational option integrating current methods mostly based on artificial neural networks. RGB-based image phenotyping can effectively predict key quality traits in olive fruit supporting the transition of the olive sector towards a digital agriculture domain.Formattato: Italiano (Italia)Formattato: Italiano (Italia) Formattato: Italiano (Italia)

    Keywords: colorimetric indexes, Hysteresis, Olea europaea L., Plantation systems, SHD, nir, RGB Tipo di carattere: Non Corsivo

    Received: 29 Jun 2024; Accepted: 23 Sep 2024.

    Copyright: © 2024 Montanaro, Carlomagno, Petrozza, Cellini, Manolikaki, Koubouris and Nuzzo. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

    * Correspondence: Giuseppe Montanaro, University of Basilicata, Potenza, Italy

    Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.