Artificial intelligence models to support the sustainability of oil palm
Dr Bonet is working with colleagues at De Montfort University to develop artificial intelligence (AI) models to classify the maturity of oil palm fruit. Correctly assessing the maturity of the fruit has a major impact on the quality and quantity of the oil produced, reducing wastage, and maximising yield and resource efficiency. The low-resource AI model approach will broaden its potential use and benefit growers, large producers and small farmers, supporting the Academy’s strategic goal of building a sustainable society.
Judging maturity in oil palms
Palm oil is a highly versatile and widely consumed vegetable oil. Today it is in great demand worldwide because of its many uses: food, cosmetics, production of paints, coatings, adhesives, lubricants, and as a raw material in the production of biodiesel. Palm oil also plays a fundamental role in reducing poverty in the areas where it grows by employing millions of people in the sector.
Fruit maturity directly affects the yield and quality of the palm oil produced. As the fruit matures, the amount of oil in the fruit increases, resulting in a higher oil yield. By classifying the maturity of the fruit, it is possible to optimise it for harvesting and maximise oil production. Oil extracted from riper fruits tends to have a more desirable fatty acid profile, with lower free fatty acid content and better oxidative stability. Therefore, the maturity classification helps to obtain better quality oil.
There are guidelines used to train palm collectors in how and when to collect the palm fruit, but these can be difficult to apply. Having a tool to help in this is key to efficient and sustainable production. It will help reduce wastage by improving the planning and scheduling of the oil palm harvest by avoiding over-ripeness or loss of fruits.
AI tools to support farmers
Dr Bonet has applied her research experience in the AI and big data field to agriculture in this project, using AI methods to create a low-resource model to classify the quality of fruit maturity. The DIA-supported project works with researchers at De Montfort University in the UK and UNIPALMAS, a palm growing company that has provided the necessary data to be able to develop the models.
“I am in a project with professors from De Montfort University and I saw it [the DIA programme] as an opportunity to strengthen this alliance and deepen the project.”
As well as fostering collaborations with UK partners and a range of producers, the project aims to contribute to economic inclusion and a sustainable society, supporting the Academy’s strategic goal. By providing clear guidelines and tools to assess fruit maturity, the project enables small farmers to optimise harvest time, increase oil production, and improve their income. This contributes to reducing income disparities and promoting economic inclusion within the palm oil industry. Accurate fruit maturity grading also helps farmers optimise their harvest schedules, reducing waste of overripe or underripe fruit, further reducing the environmental impact of oil palm cultivation and contributing to more sustainable agricultural practices.
There is also a social impact from providing tools and knowledge for fruit maturity grading. The project promotes best practices and encourages farmers to adopt sustainable approaches that minimise negative social consequences such as conflicts over land or violations of labour rights.