As its supply is destabilised by world events, palm oil’s role in meeting global demand for vegetable oil has been magnified yet again. FAO reported in 2021 that palm oil feeds 34% of the world’s vegetable oil demand with significantly less land area compared to other vegetable oils but still holds great potential for improvements in yield and quality. In fact, factors like suboptimal planting locations or poor management practices have affected the ability of oil palm to ensure the superior efficiency it promises over other oil crops – whether directly or indirectly, this has spurred more land to be made available for its cultivation and further compounds negative environmental impacts.
As the demand for sustainable palm oil grows, it becomes critical for the industry to prioritise the health of oil palm plantations to maximise yield and produce better quality oil. To support this effort, Adatos.AI has developed a sophisticated machine learning model that identifies areas suffering from various types of crop stress. Through derivative vegetation indices (VIs) produced from remote sensing, our model recognises areas within oil palm plantations that suffer from a range of ailments. A variety of VIs were used to identify different indicators of health – for instance, we capitalised on red-edge reflectances to magnify the spectral signatures from chlorophyll content which is indicative of nutrient content to spotlight deficiencies. Our model has been able to detect complex situations such as diseases arising from both flooding and drought conditions.
Insights from our model have been used to optimise plantation management and improve potential productivity and yield. Alongside other sustainable farming practices and promoting biodiversity, such efforts are aimed at helping transform oil palm agriculture into its most efficient, productive, and ultimately sustainable form.