Please login first
Leveraging Machine Learning for Early Detection and Monitoring of Xylella fastidiosa in Olive Cultivation: Implications for Technological Diffusion
, , * , *
1  Instituto de Agroecoloxía e Alimentación (IAA), Universidade de Vigo, Nutrition and Food Group (NuFoG), Campus Auga, 32004 Ourense, Spain.
Academic Editor: Milena Horvat

Abstract:

Introduction

Xylella fastidiosa subsp. pauca (Xfp) has caused significant economic losses in European agriculture. Xfp triggers olive quick decline syndrome (OQDS), disrupting the flow of ecosystems, biodiversity, and economic stability. To control Xfp, the EU has implemented quarantine protocols, including eradicating infected plants, vector suppression, and restriction of trade barriers in plant material. In this context, using Machine Learning (ML) technologies can help control pest pressures, while Deep Learning (DL) algorithms can detect spectral signatures in leaf reflectance caused by infection with over 90% accuracy.

Aim and methodology

This systematic review analyzes the application of ML for the early detection of Xfp. Academic databases (Scopus, PubMed, and ScienceDirect) were searched using keywords such as “Xylella fastidiosa,” “plant disease,” and “machine learning.” Only peer-reviewed studies published between 2015 and 2022 discussing the use of ML for Xfp or OQDS diagnosis in olives were selected. Those focused exclusively on unrelated pests or generic Artificial Intelligence (AI) applications were excluded. A total of 17 studies satisfied the inclusion criteria.

Results

ML enables an adaptive, data-driven, and efficient method of farming. The major input here is a review-based study that uses hyperspectral imagery technology and thermal data from Unmanned Aerial Vehicles and satellites. The most popular models applied are supervised ML algorithms and DL architectures. Yet, these analyses were conducted with small datasets or under controlled laboratory conditions without adequate validation at a large scale in the field, thus impacting the generalization of the results. Further barriers to entry include cost of data acquisition, farmer uptake, regulatory obstacles, and interoperability issues.

Conclusion

ML-enabled sensing technologies complement and supplement existing phytosanitary measures with precision agriculture support for crop health, water use, and Xfp detection. The success factors are early detection and strong field validation involving regulatory and farm-level decision-making frameworks.

Keywords: Xylella fastidiosa; Olive quick decline syndrome (OQDS); Olive cultivation; Machine learning; Deep Learning
Comments on this paper
Currently there are no comments available.


 
 
Top