Artificial intelligence (AI) is helping with the improving of pollination and crop yield via a new research project to track insects including honey bees, moths and hover flies.
Newly published research in the International Journal of Computer Vision, outline the recording of many pollinators at a commercial strawberry farm in Victoria, that built a database of over 2000 insect tracks.
The recordings are then analysed using a field of AI called computer vision, that helps to interpret and understand the visual world.
That AI data will be used to track individual movements of insects, count them, and to monitor their flower visits.
This enabled farmers and researchers to understand the contributions of different species to pollination.
Monash University lead researcher Malika Ratnayake said the research used specific custom algorithms researchers built.
"It works via two components, a deep learning component and a usual traditional image processing component," Dr Ratnayake said.
"So when an insect appears in the video, it uses a deep learning component to identify, detect the insect and classify it to their type."
Dr Ratnayake said the deep learning AI model detects four types of insects - honey bees, hover flies, moths and wasps and after it is identified the two simultaneous components work together to track the insect throughout frames of the video.
He said optimal pollination requires the right number of pollinator visits to flowers and the research focused on "which insects visited flowers" with honey bees visiting the most while wasps were more inclined to hunt for other insects.
He said while there is more research into the use of AI tracking that needs to be done over multiple crop cycles, the overall impacts for farmers could be beneficial in different ways.
"For farmers, they can use this kind of technology to monitor pollination in different parts of the farm and take necessary actions if there are less number of insects visiting the crops in particular part of the farm for example," Dr Ratnayake said.
"They can do some small adjustments like changing bee hive locations, the number of bee hives, or they can open and close entrances or sidewalks in greenhouses,
"Or they can look into introduce different types of flowers to attract pollinators for areas where there is a large number of insects.
Research co-author Alan Dorin, said traditional methods of insect monitoring can be time-consuming, and can give inaccurate data.
He said research conducted is a step toward correct and timely information for farmers.
"The monitoring system developed through this study can generate same-day data of crop pollination levels and provide farmers the evidence they need to inform decision-making," Associate Professor Dorin said.
"Knowing the extent to which a crop has been pollinated allows growers to alter hive locations and numbers to boost pollination levels.
"Farmers might also open or close greenhouse sidewalls to encourage or discourage insect visits from particular directions.
"They may decide to add flowers to entice insects to explore crop regions that have not been pollinated adequately.
Researchers will be collaborating and working with a number of institutions and growers in coming months, including members of the Australian Blueberry Growers Association, Costa Group's berries division, and the CSIRO.
Dr Ratnayake said there are research collaborations planned with the University of Trento, Italy and the German Centre for Integrative Biodiversity Research, Halle-Jena-Leipzig, Germany, to explore precision pollination and the use of AI in tracking European insects.