Overview

The MANIFESTO project (Machine leArNIng For intElligent inSecT mOnitoring) aims to develop an innovative AI-powered early detection system for the Asian Citrus Psyllid (ACP), a pest that transmits the devastating Huanglongbing (HLB) disease in citrus crops. The project addresses the urgent threat posed by ACP’s recent detection in Cyprus (2023) and its potential spread to other EU and Mediterranean regions. By integrating artificial intelligence (AI), remote sensing, and causal machine learning, MANIFESTO seeks to revolutionize pest management, reduce pesticide use, and safeguard Cyprus’s citrus industry (valued at €8 million annually).

Project Objectives

The MANIFESTO project aims to combat the invasive Asian Citrus Psyllid (ACP) through a four-pillar AI-driven approach. First, it will develop a benchmark dataset of ACP images to train detection models (O1), addressing the lack of Cyprus-specific data. Second, it will design and optimize deep learning architectures (e.g., YOLO, Faster R-CNN) for accurate pest identification on sticky traps (O2). Third, the project will integrate these models into an automated early-warning system with all-weather cameras and wireless alerts, validated in lab environments (O3). Finally, it will employ causal machine learning to analyze how environmental factors (temperature, rainfall, soil moisture) drive ACP population growth, enabling predictive pest management (O4). Together, these objectives aim to protect Cyprus’ €8M citrus industry, reduce pesticide use by 50% (aligned with the EU Farm to Fork Strategy), and establish a replicable framework for sustainable agriculture in Mediterranean regions.

MANIFESTO

GOALS

Develop Advanced Detection Tools

Create an AI-powered system to accurately identify Asian Citrus Psyllid (ACP) on yellow sticky traps, reducing reliance on manual inspection while achieving over 90% detection accuracy

Protect Cyprus' Citrus Industry

Safeguard the nation's €8 million citrus export industry and 3,000 hectares of orchards from ACP infestation and Huanglongbing disease

Reduce Pesticide Use

Support the EU's Farm to Fork Strategy by enabling targeted, timely interventions that could cut pesticide use by 50% by 2030

Establish Predictive Capabilities

Use causal machine learning to identify environmental drivers (temperature, rainfall, vegetation health) of ACP population growth for proactive pest management

Create Transferable Knowledge

Generate 3+ peer-reviewed publications and develop methodologies applicable to other Mediterranean citrus-growing regions

Build Collaborative Frameworks

Foster partnerships between researchers (ECoE), government (ARI), and industry (Phassouri Plantations) for sustainable agriculture innovation