A team of researchers from the Instituto de Investigación para la Gestión Integrada de Zonas Costeras of the Universitat Politècnica de València (UPV) in Gandia – Arman Foroughi, José Miguel Jiménez and Jaime Lloret – has designed a mobile application that enables early detection of diseases and pests in orange trees.
Can you imagine being able to tell whether your orange trees are sick just by taking a photo? That is precisely what this research team at the Instituto de Investigación para la Gestión Integrada de Zonas Costeras on the Gandia Campus has achieved.
Researchers Arman Foroughi, José Miguel Jiménez and Jaime Lloret have developed a mobile app capable of detecting diseases and pests in orange trees at an early stage, achieving an impressive 99.58% accuracy rate.
Technology serving the field
“The idea was to create an accessible tool for citrus growers to help them improve productivity and harvests,” explains Jaime Lloret, Professor at the UPV Department of Communications.
With this application, farmers can analyse the health of their trees directly from their phones and act quickly if they detect issues such as melanosis, black spot, canker or greening, among others.
The goal: to prevent the spread of disease and reduce economic losses in the citrus industry.
More practical, more accurate, and no internet required
The team began by analysing other similar tools to overcome their limitations. The result is a more efficient, lightweight and autonomous app.
- It works without an internet connection or servers.
- It requires very few device resources.
- It can be used directly from a mobile phone to analyse photos of leaves and fruit.
In addition, the app is multi-platform, running on iOS, Android, Windows, Linux, and Raspberry Pi. There is also a version for computers and large-scale plantations, which can even send automatic email reports on tree health.
Artificial intelligence that learns from the field
Behind this tool lies deep learning technology – a branch of artificial intelligence that uses neural networks to recognise patterns and make predictions.
The model was trained on over 5,000 images of oranges – both healthy and affected – and, after fine-tuning, achieved a diagnostic accuracy of 99.58%.
The app not only distinguishes oranges from other fruits but also identifies eight types of diseases or pests, both in leaves and in fruit.
Next step: drones and intelligent robots
The future of this innovation looks even more ambitious. The UPV team plans to integrate the app into agricultural robots and drones to automate diagnosis across extensive crop areas.
“We want to combine the software with irrigation, fertilisation and sensor network systems,” the researchers explain. This will allow farmers to benefit from intelligent, fully automated management of their plantations.
A significant step forward for sustainable agriculture
This app represents not only a technological breakthrough but also a move towards more efficient, sustainable and digitalised farming. Thanks to artificial intelligence, orange growers will be able to anticipate problems and better protect their crops.
Because when innovation serves the land, technology and tradition can truly bear exceptional fruit.
References:
An edge computing wireless sensor network for diagnosing orange fruit disease.