Artificial intelligence (AI) is being used in a growing number of applications for studying and improving the environment. Two of the most innovative proposals in this field have been presented this week, in parallel, as methods to facilitate research on birds and insects. In the first case, carried out by experts from the University of Alberta (Canada), AI techniques are being used to study bird nests with computer analysis of images captured in the natural environment. On the other hand, AI is also the basis of a project led by the company Capgemini that studies insects from sounds captured in nature.
Priscilla Adebanji, a computer science student at the Augustana Campus-University of Alberta (Canada), has carried out an innovative study of the application of AI in the analysis of images of blackbird or red-winged cowbird (Agelaius phoeniceus) nests. This method could be applied in the future to various species, reducing the time now spent manually reviewing videos to identify specific birds and their behavior, explains Professor Ivana Schoepf, an expert in bird parasitism at the University of Alberta.
Currently, one of the most common ways to study birds is by listening to recordings and watching videos. The sounds are not always of optimal quality and species identification is complex due to the variety of noises. Something similar happens with videos captured, for example, with automatic cameras (phototrapping). AI applications such as those now presented allow us to improve the identification of species and individuals, as well as their movements, streamlining the analysis process and reducing human work in the most repetitive parts.
The first part of this work has included the analysis of videos captured over two years in 30 nests of this species of migratory cowbird common in North America, applying computer vision and artificial intelligence.
The work led by Priscilla Adebanji has achieved an application that offers image analysis with results that determine, for example, the exact times that birds visit their nests to feed their chicks.
For their part, in completely parallel work but with some similarity, data and artificial intelligence experts from Capgemini, in collaboration with Naturalis Biodiversity Center and Amazon Web Services (AWS), have developed artificial intelligence (AI) models that help protect biodiversity by automating and improving insect identification through sound recognition.
Identifying insects by their sounds is a very complicated task but now "the specific AI solutions developed by Capgemini will allow scientists to analyze terabytes of information in just 24 hours, as they will no longer be limited by the limits of capacity, time and geography," highlights the company.
Insects are incredibly difficult to monitor, and since there are millions of species, this process cannot be reliably performed by humans. Although image recognition works for some species, it cannot successfully detect species that go unnoticed by the human eye. This is where acoustic recognition plays an invaluable role.
Along with traditional and camera-based monitoring, AI acoustic recognition will provide new insights that will identify insect species that would otherwise go unnoticed, allowing scientists to monitor and conserve these vital populations.
Elaine van Ommen Kloeke, Program Director at Naturalis, recalls that insects play a key role in the protection of biodiversity and are of great importance for ecosystems and the pollination of crops of great value for human consumption. "The good news is that, through collaboration, facilitated by organizations like Capgemini and technological advances, we can address this global problem and protect an animal group so important to our ecosystem. With this initiative, we aim to make this technology available to research communities around the world to help them identify insects in any environment,” says Elaine van Ommen Kloeke.
