The AI model creates a two-way conversation between farmers and supermarkets where both parties would mutually benefit.
Two students of Mohamed Bin Zayed University of Artificial Intelligence (MBZUAI) have created an algorithm that could revolutionise the UAE’s agriculture industry while reducing food waste and enhancing food security.
Sarah Al Barri and Mugariya Farooq have developed a machine learning framework to detect plant disease by taking pictures using a drone and predicting crop yield through key information fed on a mobile app. The AI-powered solution for the agriculture industry consists of two major parts that work together, said the students pursuing masters in machine learning.
“In the first part, computer vision is trained to diagnose early-stage diseases in plant crops. Due to their popularity and thus impact, we chose tomatoes, potatoes and peppers. In the second part, a machine learning model is trained to precisely predict the actual yield for a planted farm. Early disease detection reduces the resources such as water, fertilisers and pesticides that could be wasted on growing diseased plants,” Al Barri said.
Early-stage detection of diseases and knowledge of the yield will help the farming community stabilise the supply and demand of different crops.
“It can enable farmers to know beforehand the amount they should produce thus they can allocate and manage their resources accordingly. If demand and supply are equal, then prices will stabilise. If fruit and vegetable prices do not fluctuate, supermarkets will not have to dispose of the huge quantities of food they normally do, thereby limiting food wastage,” Farooq noted.
The AI model creates a two-way conversation between farmers and supermarkets where both parties would mutually benefit.
“Using AI, farmers would be able to have informed and planned financial decisions about which plant to eliminate early, and the amount of resources they should invest in planting new crops,” Farooq underlined.
Explaining how AI can be used in agriculture, Farooq highlighted the role of drones in the project.
“AI models need to be incorporated in an infrastructure or a tool to be used. Although not applied yet, our suggestion and vision for our AI model is that the plant disease detection task could be implemented on a drone that takes pictures, processes them, and provides a full diagnostic report. For the crop yield prediction task, a mobile app is to be developed where a farmer could provide information about the type of crop, water input, fertiliser and pesticide use, and the land area, and the output of the app gives the predicted yield per area.”
The duo made the sustainable solution for a two-day agricultural hackathon held by the Abu Dhabi Agricultural and Food Security Authority (ADAFSA) in November.
“The hackathon had two tracks: biosecurity preparedness analysis and food security. We chose food security. Our decision was motivated by powerful statistics. For example, in the UAE, an average person wastes around 224 kg of food each year, and we believed an AI-powered solution would be capable of tackling this problem,” Farooq added.
The project won second place at the agricultural hackathon and bagged an $8,167 prize. They have featured the solution at the Masdar Innovate stand during the Abu Dhabi Sustainability Week and the World Future Energy Summit. More recently, at the UAE Innovates event in Expo 2020 Dubai.
Earlier this month, Ne’ma – a new national initiative to address the issue of food loss and waste, was launched in the country. It aims to encourage public and private sector entities to collectively address food waste and encourage responsible consumption to preserve food resources for a sustainable future.
