YIELD Test – Yield Improvement using ELectrophysiology Device

• Forschungsstelle: BLW
• Projektnummer: 19.27

Schlüsselwörter (Deutsch)

Ertrag, Nährstoffmangel, Elektrophysiologie, Gartenbau, Maschinelles Lernen

Schlüsselwörter (Englisch)

yield, nutrient deficiency, electrophysiology, horticulture, machine learning

Schlüsselwörter (Französisch)

rendement, carence en nutriments, électrophysiologie, horticulture, apprentissage machine

Kurzbeschreibung (Englisch)

This digital agriculture project delivers benefits for Swiss growers:
• reduced costs (pesticides, water, fertilizers and labor)
• reduced losses to diseases and blights
• evidence-based plant physiology data for customers and regulators
• recorded insights of skilled growers
• improved scouting effectiveness

This project uses electrophysiology for early diagnosis of stress in greenhouse tomatoes and other crops. Following successful research at Agroscope, a system for interpreting small electrical signals, inherent in all higher plants has been developed and accurately shows (>90%) if a plant is stressed due to water or nutrient deficiencies or insect infestations. Diagnoses are made several days before visible symptoms are evident. Early warning of stress enables producers to optimize growing conditions, apply effective and environmentally preferable crop protection strategies and increase yields.

Three research questions will be tackled by a multi-disciplinary team:
1. can algorithms diagnosing stress in tomatoes be applied to other crops
2. can specific algorithms for crop stress from drought, nutrient deficiencies and insect attacks be combined to create a generalized measure of plant stress
3. can electrophysiology measurements be used to enhance scouting in commercial greenhouses.

Experimental work will be done at Agroscope Conthey and verified at Domaine des Mattines, a leading Swiss grower. This project enhances Switzerland’s leadership in applied plant electrophysiology.

Kurzbeschreibung (Französisch)

Ce projet d'agriculture numérique présente des avantages pour les producteurs suisses :
- réduction des coûts (pesticides, eau, engrais et main d'œuvre)
- la réduction des pertes causées par les maladies et le mildiou
- des données probantes sur la physiologie des plantes à l'intention des clients et des organismes de réglementation
- des renseignements consignés sur les cultivateurs qualifiés
- amélioration de l'efficacité du repérage des facteurs de stress

Ce projet utilise l'électrophysiologie pour le diagnostic précoce du stress chez les tomates de serre et d'autres cultures. Suivant le succès de la recherche chez Agroscope, un système d'interprétation des petits signaux électriques, inhérents à toutes les plantes hautes, a été développé et montre avec précision (>90%) si une plante est stressée par des facteurs comme : le manque d'eau et de nutriments ou l'infestation d'insectes. Le diagnostic est posé plusieurs jours avant l'apparition des symptômes visibles. L'alerte précoce du stress permet aux producteurs d'optimiser les conditions de croissance, d'appliquer des stratégies de protection des cultures efficaces et respectueuses de l'environnement ainsi que d'augmenter les rendements.

Trois questions de recherche seront abordées par une équipe multidisciplinaire :
1. Les algorithmes de diagnostic du stress chez la tomate peuvent-ils être appliqués à d'autres cultures ?
2. Les algorithmes spécifiques au stress des cultures dû à la sécheresse, aux carences en nutriments et aux attaques d'insectes peuvent-ils être combinés pour créer une mesure généralisée du stress des plantes ?
3. Les mesures électrophysiologiques peuvent-elles être utilisées pour améliorer le repérage des facteurs de stress dans les serres commerciales ?

Des travaux expérimentaux seront effectués à l'Agroscope de Conthey et vérifiés au Domaine des Mattines, un leader dans le mairaichage Suisse. Ce projet renforce le leadership de la Suisse en électrophysiologie végétale appliquée.

Projektziele (Englisch)

This project tests three key hypotheses.
1. It is possible to produce a useful algorithm which provides a generalized alarm for plant stress in advance of visual symptoms using plant electrophysiology. This general algorithm will be produced by reviewing commonalities in existing and proven specific algorithms and will be refined and tested by experimentation.
2. Both the specific algorithms of plant stress developed using tomatoes and the general stress algorithm can be applied on a wider range of crops. Algorithms developed for tomatoes will be tested on other crops, initially on greenhouse crops and later in field crops. The hypotheses will be tested using experimentation and rigorous statistical analysis.
3. Electrophysiological signals, which are interpreted using machine learning, can be used to improved yields in commercial tomato operations. Information can be applied in different use-cases in a commercial operation:
a. Continuous detect and alert monitoring at fixed locations reducing the frequency of scouting.
b. Incorporation in automated control systems, for example to control nutrigation.
c. As a scouting tool in which skilled operators combine their own experience and observations with electrophysiology data from a portable device.

Algorithms will be tested to see if they detect plant stress in advance of visual symptoms, at the same time or later than a skilled plant scout. This work will be carried out at Agroscope in conjunction with the above experiments and then the most promising algorithms will be tested at Domaine des Mattines. Key data to be collected include water and nutrigation inputs, climate and environmental conditions, plant growth and yields, physiological disorders on fruits/leaves, plant electrophysiology.

Abstract (Englisch)

YIELD Test showed that crop electrophysiology monitoring shows strong value as an agronomic tool, specifically to:

1. Diagnose crop stress in advance of visual symptoms. General stress models were developed and deployed showing high accuracies and strong grower engagement. Growers are already benefiting from early stress diagnosis and a scientific publication has been published.

2. Improve greenhouse management. Autonomous irrigation trials were successfully completed, although further refinements are needed prior to commercialization, showing strong promise for reducing labor required to manage irrigation.

In addition, a method was developed to assess the universality of crop signals to understand how easily and widely crop electrophysiology monitoring could be deployed. A cross-validation approach using tomatoes and eggplants shows good results with enough universality between closely related species for easy deployment of electrophysiology tools. The method will be applied to a wider range of species to draw clear conclusions about signal universality.

Plant electrophysiology provides early warning of crop stress and routes towards data driven and autonomous growing which will reduce costs and crop inputs. Project results were shared in face-to-face meetings with growers, via academic publications and conferences, in horticulture trade press and via social media. Switzerland’s position as the leader in applied plant electrophysiology was reinforced.