Artificial intelligence can help create the pollen jigsaw of current and ancient plants

An emerging system that combines rapid imaging with artificial intelligence can help scientists build a comprehensive picture of current and historical environmental changes by quickly and accurately analyzing pollen grains.

Pollen grains of different plant species are unique and identifiable by their shape. Analyzing which pollen grains are captured in samples such as lake sediment cores helps scientists understand which plants were thriving at any point in history, potentially thousands to millions of years old.

Until now, scientists have measured pollen types in sediment or air samples by hand using a light microscope—a specialized and time-consuming task.

Now, scientists from the University of Exeter and Swansea University are combining cutting-edge technology with imaging flow cytometry and artificial intelligence to make pollen identification and classification even faster.

In addition to creating a complete picture of past flora, the team hopes the technology can one day be applied to more accurate pollen readings in today’s environment, which could help those with hay fever to reduce symptoms. The paper is titled “Deductive Automated Pollen Classification in Environmental Samples Using Exploratory Deep Learning and Imaging Flow Cytometry”, and it is published. New Phytologist.

Exeter University Dr. Ann Power said, “Pollen is an important ecological indicator, and piecing together the jigsaw of different pollinators in today’s and past environments can help us build a picture of biodiversity and climate change.”

“However, identifying which plant species pollen belongs to under a microscope is incredibly labor-intensive and cannot always be done. The system we are developing will reduce this time and improve classification. This means we can build a richer picture. Pollen in the atmosphere More rapidly, it reveals how climate, human activity and biodiversity have changed over time, or better understanding what allergens are in the air we breathe.

The team has already used the system to automatically analyze slices of a 5,500-year-old lake sediment core, rapidly classifying more than a thousand pollen grains. In the past, it took experts eight hours to count and sort—a task the new system completed in under an hour.

The new system uses imaging flow cytometry — a technology typically used in medical research to examine cells, to quickly capture pollen images. A unique type of artificial intelligence based on deep learning has been developed to identify different types of pollen in an environmental sample. It is capable of distinguishing even when the sample is incomplete.

Swansea University Dr. Claire Barnes said, “Until now, AI systems developed to classify pollen grains learn from and test the same pollen library – meaning each sample is perfect and matches the species the network has seen before. These systems cannot identify pollinators in the environment with a few bumps along the way or training. Pollen not included in the library cannot be classified.”

“Incorporating a unique version of deep learning into our system means that artificial intelligence is smarter and applies a more flexible approach to learning. It can deal with poor quality images and use shared species characteristics to predict which family a pollinator belongs to. If during training The system may not have seen it before.”

In the coming years, the team hopes to refine and launch the new system and use it to learn more about grass pollen, a particular nuisance for hay fever sufferers. Dr. Power said, “Some grass pollens are more allergenic than others. If we can better understand which pollens are prevalent at certain times, that will improve pollen forecasting to help people with hay fever reduce their exposure.”