Insect Wingbeats Can assist in Increasing Biodiversity

Insect populations are collapsing around the world, with serious consequences for our ecosystems and no one knows why. A new AI method developed by the University of Copenhagen is set to aid in the monitoring and cataloguing of insect biodiversity, which has previously been difficult.

Insects play critical roles in nature as plant pollinators, food sources for a wide range of animals, and decomposers of dead matter. They have, however, struggled in recent decades. It is estimated that 40% of insect species are declining, and a third are endangered.

As a result, it is more important than ever to monitor insect biodiversity in order to understand their decline and, hopefully, assist them. So far, this task has been difficult and time-consuming. This is due, in part, to the fact that insects are small and highly mobile. In addition, scientific researchers and government agencies must set up traps, capture insects, and study them under a microscope.

To overcome these challenges, University of Copenhagen researchers developed a method that recognizes and detects individual insect wingbeats using data from an infrared sensor. The AI method is based on unsupervised machine learning, which allows the algorithms to group insects of the same species without the need for human intervention.

Use of Artificial Intelligence

The researchers have already created an algorithm for detecting pests in agricultural fields. However, rather than identifying insects as pests, the researchers were able to develop this new algorithm to identify and count various insect populations in nature based on sensor measurements.

"The sensor is similar to wildlife surveillance cameras, which are used to monitor the movements of larger animals in the wild. Instead of taking a picture, the sensor counts the number of insects that have flown into the light source. The algorithm then divides the insects into groups based on their wingbeats” Assistant Professor Raghavendra Selvan of the Department of Computer Science, who led the development of the sensor's artificial intelligence, explains.

The algorithm distinguishes insects by their silhouettes when their wings are folded out because their physical differences are most visible when their wings are folded out. It then compares the silhouettes of different insect recordings and groups similar silhouettes together, which can then be used to determine which insect flew through the light beam.

Prototype to be Released in Spring

When the insects begin to emerge in full force in the spring, scientists will use the initial prototype to go out into nature and collect real-world data.

Until now, researchers have tested the algorithm and artificial intelligence on a large image database of insect recordings obtained under controlled conditions as well as some real-world data, with promising results.

"We will put the sensor to the test in a variety of environments, including heathland, forests, and agricultural areas, to see how it performs in the real world. However, in order to feed the algorithm more data and make it more accurate,” says Raghavendra Selvan.

According to the researchers, their invention allows them to thoroughly monitor many geographical areas than was previously possible. At the same time, the invention reduces the number of resources required to keep an eye on insects, which account for 80 percent of all terrestrial animal species.

"It is currently impossible to afford the level of monitoring required to gain a more precise picture of how our insects are faring. This sensor only requires humans to deploy it in the wild. Once there, it begins gathering information on local insect populations," concludes Klas Rydhmer.

(Source: University of Copenhagen)