This article has been reviewed according to the editorial process and policies of Science X. The editors have emphasized the following features while ensuring the credibility of the content:

fact checked

peer reviewed publication

reliable source

proofreading






In a new study, a type of artificial intelligence called “computer vision” has been taught to identify the components of a plant in the same way a human would. Credit: Shutterstock/Pathomrat Praerin

Researchers from UNSW and the Sydney Botanic Gardens have trained artificial intelligence to unlock data from millions of plant specimens stored in botanical gardens around the world to study and combat the effects of climate change on vegetation.

“Herbarium collections are incredible time capsules of plant specimens,” says the study’s lead author, Associate Professor Will Cornwell. “Every year over 8,000 specimens are added to the National Herbarium of New South Wales alone, so it’s no longer possible to go through things manually.

Using a new mechanical algorithm to process more than 3,000 leaf samples, the team discovered that, contrary to the common pattern between species, leaf size does not increase in warmer climates within a single species.

Published in American Journal of Botanythis research not only shows that factors other than climate have a large effect on leaf size within a plant species, but also demonstrates how artificial intelligence can be used to transform static collections and to document the effects of climate change quickly and efficiently.

Herbarium collections move into the digital world

Herbaria are scientific collections of plant specimens that have existed since at least the 16th century. “Historically, it was a valuable scientific effort to go out, collect plants and then store them in a herbarium. Every record has its time and place and collector and potential species identifier,” says A/Prof. Cornwell, researcher at the School of BEES and member of the UNSW Data Science Hub.

A few years ago, to facilitate scientific collaboration, there was a movement to bring these collections online.

“The botanical collections were locked in small boxes in specific locations, but the world is very digital now. So to bring information about all the amazing specimens to scientists who are now scattered around the world, efforts were made to scan specimens to produce high-resolution digital copies of them .”

The largest graphics project was undertaken at the Sydney Botanic Gardens when over 1 million plant specimens at the National Herbarium in New South Wales were converted into high resolution digital images.

“The digitization project took over two years, and shortly after it was completed, one of the researchers, Dr. Jason Bragg, contacted me from the Sydney Botanic Gardens. He wanted to see how we could integrate machine learning with some of these digital high-resolution images of Herbarium specimens. ”

Dr. Bragg says, “I was excited to work with A/Prof. Cornwell to develop models to detect leaves in plant images and then use these large datasets to study the relationship between leaf size and climate.”






The machine learning algorithm developed by the research team measures and identifies plant samples. Credit: University of New South Wales

‘Computer vision’ measures leaf sizes

Together with Dr. Bragg at the Sydney Botanic Gardens and UNSW Honors student Brendan Wilde, A/Prof. Cornwell created an algorithm that could be automated to analyze and measure the leaf size of scanned botanical samples for two plant genera – Syzygium (commonly known as lillipillies, brush cherries or satinas) and Ficus (a genus of about 850 species of woody trees, shrubs and vines).

“This type of artificial intelligence is called a convolutional neural network, also known as Computer Vision,” says Cornwell. The process essentially teaches the AI ​​to see and identify parts of the plant the same way a human would.

“We had to create a training data set to teach the computer: this is a leaf, this is a stem, this is a flower,” says Cornwell. “So we basically taught the computer to position the leaves and then measure their size.”

“Measuring the size of leaves is not innovative, because many people have done this. But the speed with which these samples can be processed and their characteristics recorded is a new development.”

Breach of frequently seen patterns

A general rule of thumb in the botanical world is that in wetter climates, such as tropical rainforests, plant leaves are larger compared to drier climates, such as deserts.

“And there’s a very consistent pattern that we see in leaves across species around the world,” says Cornwell. “The first test we did was to see if we could reconstruct these relationships from machine-learned data, which we could. But the second question was, because we now have so much more data than we had before, do we see the same thing within species?”

The machine learning algorithm was developed, validated and applied to analyze the relationship between leaf size and climate within and among species for Syzygium and Ficus plants.

The results from this test were surprising – the team discovered that while this pattern is seen between different plant species, the same correlation is not seen within a single species around the world, probably because another process, known as gene flow, is operating within species. That process weakens plant adaptation at the local scale and may prevent the leaf size–climate relationship from evolving within species.

Using artificial intelligence to predict future responses to climate change

The machine learning approach used here to detect and measure leaves, although not pixel perfect, provided an accuracy suitable for investigating relationships between leaf characteristics and climate.

“But because the world is changing quite rapidly and there’s so much data, these types of machine learning methods can be used to document the effects of climate change effectively,” says Cornwell.

What’s more, machine learning algorithms can be trained to identify trends that might not be immediately obvious to researchers. This could lead to new insights into plant evolution and adaptation, as well as predictions of how plants might respond to the future impacts of climate change.

More information:
Brendan C. Wilde et al., Analyzing trait-climate relationships within and among taxa using machine learning and botanical sampling, American Journal of Botany (2023). DOI: 10.1002/ajb2.16167

Diary information:
American Journal of Botany

#Artificial #intelligence #uncovers #hidden #features #planets #flora #save #species