Scientists build low-cost AI cameras to protect BumblebeesScientists at Oregon State University have developed a low-cost, AI-powered camera system that is capable of monitoring bumblebee populations without relying on lethal traps or expensive, manual field surveys. The research team’s findings, which were published in the journal Remote Sensing in Ecology and Conservation, show that the non-invasive technology could transform insect conservation and agricultural planning at a critical moment. Several bumblebee species face severe population declines and are being evaluated for federal protection under the Endangered Species Act.Technology overcomes traditional survey challengesHistorically, tracking pollinator populations required labour-intensive hand-netting or lethal blue-vane traps. While effective at documenting species diversity, traditional traps kill the insects being studied and direct manual observation requires extensive time and financial resources. To solve this, researchers built an affordable camera setup using off-the-shelf commercial components and tested it in a flowering red clover field at Oregon State’s Hyslop Field Research Laboratory in Corvallis.To lure daylight-active (diurnal) insects like bumblebees toward the lenses, the team tested various visual backgrounds. They discovered that placing a bullseye pattern behind the cameras drew significantly more bee visits compared to plain, solid backgrounds.Machine learning matches traditional accuracyTo process thousands of photos automatically, the researchers trained custom deep-learning computer vision models. They found that “tiled" AI models – which slice images into smaller grids to evaluate localized details – outperformed standard models analysing full-frame photos.According to the research results, during field trials, the camera system successfully documented six distinct bumblebee species. The data mirrored the species diversity captured through traditional netting and lethal trapping, proving that remote camera monitoring can deliver accurate data without harming local ecosystems.“Insects are vitally important, and we need methods to better understand their populations. This tool should help us to do that in a low-cost, easily scalable manner,” said Michael Getz, a data scientist at the Biodiversity Research Institute in Maine, who led the project as a master’s student at Oregon State University.The ability to collect large volumes of field data could prove crucial for wildlife agencies and agricultural producers. Many commercial crops rely heavily on native bees for pollination, yet conservationists often lack the widespread population data needed to implement targeted recovery plans.Andony Melathopoulos, an associate professor of pollinator health at Oregon State and co-author of the study, noted that missing data frequently complicates federal decisions regarding endangered species listings and critical habitat designations.“This technology opens up the possibility for much richer data so that we can be more precise when it comes to decisions on listing a species and characterizing critical habitat,” Melathopoulos explained.