At our insect collection we are using AI to help detect the Brown Marmorated Stink Bug, a priority insect pest.

The Brown Marmorated Stink Bug (BMSB) is a sap sucking bug native to China. It is a threat to crops such as apples, stone fruits, hazelnuts and grains.

BMSB looks similar to many other stink bug species, making it difficult to recognise.

We are working closely with biosecurity staff at the Department of Agriculture, Water and the Environment (DAWE) to help keep this pest out of Australia.

An AI app for stink bugs

Our Australian National Insect Collection has thousands of specimens of expertly identified stink bugs species from Australia and overseas.

“We are taking detailed digital images of the stink bugs in our insect collection,” our botanist Dr Alexander Schmidt-Lebuhn said.

“We’re training Artificial Intelligence to recognise BMSB and tell it apart from similar looking species, especially native ones that are commonly found by biosecurity officers.”

The result is an app that biosecurity officers are currently trialling at ports and airports. It uses the video feed of a smartphone.

“You hold the camera over the bug and can zoom in or out and look at it from different angles. The AI model in the app provides a constantly updated identification estimate and its degree of certainty,” Alexander said.

The app also has species’ profiles with example images and species information. Users can record a photo of the bug, its identification, and the geographic coordinates and local time.

A woman looking at images of the stink bug on a computer.
We are using a 3D imaging system to take images of stink bugs from many angles to train the AI inside our app. This requires several hundred images per species. 

A rapid spread around the world

Australian has about 600 named native stink bug species, as well as several thousand more undescribed species. They get their name from the smell of their pheromones, the chemicals they use to communicate.

Michael Elias is a taxonomist at our insect collection.

“BMSB began to spread in the mid-1990s when exports from China increased. It’s now established in many countries around the world,” Michael said.

“This is thanks to their biology. They gather in large numbers near bright lights to breed, then go into hibernation in dark, hidden places. If they breed near the bright lights of a car plant, the pregnant females can hide in the cars and wake months later in a new country, ready to cause an infestation.

“Unlike our native stink bugs, BMSB has no specialised natural enemies here to keep its populations in check,” he said.

A man looking at pinned specimens in a lab.
The app shows how likely it is that the species identification is correct. 

From daisies to stink bugs

Alexander is a botanist who researches the 25,000 different species in the daisy family.

“I got involved in biosecurity work because the daisy family has many weeds. This is because their seeds spread very easily,” he said.

“We developed an app prototype for identifying the seeds of weedy daisies in collaboration with Microsoft. We’ve used what we learned there to develop our stink bug app.

“We hope to expand it in two different directions. Firstly, we want to add AI models for more types of biosecurity threats, beyond stink bugs. Secondly, we hope to involve the public in biosecurity work so that people can identify and report pests and weeds.”

1 comments

  1. Extrapolating thru a wishful lens, I can see this technology installed in an autonomous robot that roams up and down orchard rows, seeking pests and applying a directly targeted dose of species specific pesticide, or maybe a jet of steam or liquid nitrogen. Could be the way to go in circumstances that don’t lend to alternative pest control systems.

What do you think?

We love hearing from you, but we have a few guidelines.