Researchers from the University of Edinburgh have developed a machine learning-powered imaging system that enables farmers to identify which crops thrive in poor soil conditions.
Working closely with colleagues from Ethiopia, researchers tested the new imaging system on chickpeas; a popular food crop which provides a primary source of protein for millions of people.
The research, published in The Plant Journal, was carried out in collaboration with the Ethiopian Institute for Agricultural Research and Addis Ababa University.
As part of the study, tall, transparent containers – known as rhizoboxes – were used to grow individual chickpea plants and to study their roots.
Controlled by Raspberry Pi, a photographic imaging station which used a network of simple cameras was then deployed to capture the growth of the root systems over time.
The research team also developed machine learning-powered software to collect and analyse the imaging data and fill in the gaps where roots were concealed by soil. This system also includes a frame that holds several rhizoboxes, allowing for the simultaneous testing of different soil conditions within separate boxes
Long-term, researchers said the new approach could help institutions across the world improve their plant breeding programmes and provide vital food sources for millions.
Dr Peter Doerner, Personal Chair of Applied Biology at the University’s Institute of Molecular Plant Science, said the new techniques could also boost crop resistance.
“This new approach opens up huge opportunities for increasing crop resilience against soil and climate challenges across the globe by enabling crop breeding for better resource capture by roots,” he said.
“It could also be used to develop more resilient UK crops, such as barley used in the whisky industry, which suffer due to increasingly common dry spells in early spring,” Dr Doerner added.
Previous studies have highlighted that the size, depth and spread of root systems have an enormous impact on how effectively plants can absorb vital nutrients and tolerate challenging conditions, such as drought.
Currently, instruments which examine root growth exist. However, researchers noted that they are often costly and complex devices which farmers and plant breeders in low-income countries cannot afford.
Similarly, it can also take several years to test specific crops and identify which varieties are best adapted to local soils and environmental conditions.
- New AI-based test could identify Covid-19 within an hour
- Exciting 5G projects roll-out in six cities across Britain
- Australia to force tech giants to pay for news in landmark move
Dr Solomon Chanyalew, Director at the Debre Zeit Research Centre, Ethiopia, said the imaging methods are also applicable to other crops and are highly adaptable based on soil conditions.
“This is a technically simple and affordable tool and technique with profound effect on root system studies,” he said.
“We have seen impressive results in the chickpea rhizobox technology and feel it would be effectively used in other legumes, like lentil beans too, and different soils.”