Optimizing Precision Agriculture Artificial Intelligence (AI) Models for Improved Performance and Broader Applications

Current practices in precision agriculture often rely on disjointed data streams and static models, limiting their effectiveness in detecting and responding to emerging plant health threats. One of the most pressing challenges is timely and accurate plant disease recognition, which is critical for reducing crop loss, improving yield, and minimising unnecessary chemical use. This project addresses these challenges by developing advanced AI-based solutions focused specifically on plant disease detection and management.

The key challenge lies in creating intelligent models that can integrate diverse data sources, including high/low resolution leaf imagery, for early and accurate diagnosis of plant diseases. Existing systems often lack adaptability across different crop types and growing conditions, resulting in delayed or inaccurate responses to disease outbreaks.

To address these limitations, the project will develop and validate novel AI models that utilise advanced neural networks, particularly in image analysis and spatiotemporal data fusion, to improve the accuracy and speed of plant disease recognition. These models will be optimised for real-time application in field conditions, enabling both smallholder and industrial-scale farmers to make informed decisions on treatment and prevention strategies.

Deliverables:
AI Models: Development of image-based and multimodal AI models for early detection and classification of plant diseases, trained on diverse datasets representing a range of crops and environmental scenarios.
Publications: Three peer-reviewed journals or conference papers documenting the experimental results and practical recommendations for adoption in diverse agricultural contexts.