Accurate models for early harvest estimation in citrus production generally involve expensive variables. The goal of this research work was to develop a model to provide early and accurate estimations of harvest using low-cost features. Given the original data may derive from tree measurements, meteorological stations, or satellites, they have varied costs. The studied orchards included tangerines (Citrus reticulata x C. sinensis) and sweet oranges (C. sinensis) located in northeastern Argentina. Machine learning methods combined with different datasets were tested to obtain the most accurate harvest estimation. The final model is based on support vector machines with low-cost variables like species, age, irrigation, red and near-infrared reflectance in February and December, NDVI in December, rain during ripening, and humidity during fruit growth.
Highlights:
Red and near-infrared reflectance in February and December are helpful values to predict orange harvest.
SVM is an efficient method to predict harvest.
A ranking method to A ranking-based method has been developed to identify the variables that best predict orange production.
In order to define management and marketing strategies, farmers need adequate knowledge about future yield with the greatest possible accuracy and anticipation. In citrus orchards, greater variability and non-normality of yield distributions complicate the early estimation of fruit production. This study was conducted with the objective of developing a method to estimate citrus load based on orchard characteristics, morphological information of trees and number of fruits in defined locations of the crow. Field data from 16 citrus orchards obtained from 2005/06 through 2013/14 seasons were used. Machine learning techniques were applied to predict yield; these methods can reduce the estimation error as well as decrease the need for in-field measuring, thus reducing both the cost and time of the process.
El objetivo de este trabajo fue modelar el crecimiento de limonero ‘Eureka’ (Citrus limón (L.) Burn f) injertado sobre Limón Rugoso (Citrus jambhiri Lush), desde los 60 días después de plena floración hasta la cosecha, en plantaciones comerciales, de la provincia de Corrientes, Argentina. Para ello se registró el diámetro ecuatorial de frutos desde los 60 días después de plena floración hasta el momento de cosecha en cuatro ambientes de la provincia de Corrientes, Argentina. Para describir el crecimiento se compararon los modelos no lineales de tipo sigmoideo: logístico, Gompertz y monomolecular. Se utilizaron como criterios de selección de los modelos: AIC, BIC y el CME (cuadrado medio del error). La precisión del modelo seleccionado se obtuvo mediante validación cruzada. Los modelos seleccionados para describir el crecimiento de frutos de limonero ‘Eureka’ fueron los modelos monomolecular y logístico, los cuales tuvieron un error de estimación que fue medido en términos del CME con valores menores a 20,03. Los parámetros estimados para los modelos seleccionados resultaron significativos (p < 0,01). De esta manera, productores, y empresas exportadoras citrícolas de la región podrán contar con una herramienta con la que puedan predecir la producción al momento de cosecha y así para prever estrategias de comercialización.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.