B.C. researchers developing ‘insect forecasting’ system

Researchers at the University of British Columbia are piloting new technology that could one day predict the level of insects similar to the way the weather is forecast.

Over the summer, the team is deploying about 60 special sticky traps that make use of cameras, small computers and artificial intelligence to collect real time data about when bugs are around.

“Generally when you trap insects you have to manually go there and harvest (from) traps once a week and then you know this week you had (a) specific insect (count),” said Quentin Geissmann, postdoctoral research fellow in microbiology and immunology in the faculty of land and food systems, and co-lead of the Sticky Pi project.

“You don’t know when in the week the insects got trapped, and it takes a tremendous amount of time to actually ID the insects.”

The system works by taking photos every 20 minutes, including at night, then using machine learning to process the images, and more quickly identify the insect species than a human could.

Geissmann said the technology could eventually provide important information on insect behaviours for use in climate change research, such as how they respond to warming daytime temperatures.

But the tools could also have important practical applications for agriculture, he said, allowing farmers to target their pest response in real time.

0:51 B.C. government applies to spray for a dangerous invasive insect

The system has already allowed the team to identify the arrival of an invasive strawberry-eating weevil, he said.

“In the context of controlling for pests for instance — we have a lot of agricultural pests — if you are going to do precision agriculture you want to target insects when they are actually around,”  he explained.

For the time being, the project is funded solely for deployment at UBC’s Point Grey campus, as a part of the university’s Campus as a Living Laboratory program.

However the team hopes to deploy the system across the City of Vancouver in 2023, and eventually further beyond.

Geissmann said all of the components of the system are open source, allowing people to assemble their own traps using 3D printed components and cheap and easy to program Raspberry Pi computers.

“We’re not keeping it for us, we’re giving the tools to everyone so they can build it,” he said.

“It’s like if you are (establishing) weather stations. We want to put them kind of everywhere to make a map of where insects are and when. The ambition is to go even beyond Vancouver.”

You can find out more about the Sticky Pi project on its website.