James Cook University researchers have developed a focused tutorial to help advance the use of artificial intelligence in agriculture.
The study explores multimodal AI systems that integrate text, imagery, and sensor data to provide actionable insights, enabling farmers and agribusinesses to use multimodal AI for enhanced decision-making in the field.
These systems can process satellite and field imagery, soil data, and weather information within a single platform to support farmers in making informed decisions.
Lead author and JCU master's student Mohammadreza Haghighat explained that the guide details how multimodal AI functions, its current applications in agriculture, and how researchers for smart agriculture can adapt these systems for their specific needs.
"Our aim was to create a unified framework for AI researchers to apply multimodal language models in agriculture," Haghighat said.
"We're exploring how models like ChatGPT, which process multimodal data, can be tailored for farming applications."
By feeding agriculture-specific data into these models, the research demonstrates how to generate meaningful outputs, such as identifying crop issues from drone imagery and plant health monitoring to recommend precise interventions.
This approach can optimise resource use, save time, and address other agricultural challenges.
"The guide emphasizes developing state-of-the-art AI tools more accessible to farmers," Haghighat said.
"We're working toward integrating these models into practical tools like drones, ensuring they are tested in agricultural contexts rather than purely technical domains."
The research highlights that open-source tools and affordable hardware are making multimodal AI increasingly accessible, enabling farmers to adopt these technologies for improved decision-making in the near future.
