Development and Evaluation of a Low‐Cost, Semi‐Automated Camera Trap for Surveying Bumble Bee Communities

Widespread declines in insect diversity and abundance underscore an urgent need for standardized, nonlethal monitoring methods for important pollinators such as bumble bees ( Bombus spp.). Camera traps are widely used for nonintrusive, continuous surveys of large animals but have not yet been extensively adopted for monitoring diurnal insects. Automated camera traps could improve current insect survey methods, which often rely on lethal traps or in‐person observation. Here we developed an open‐source, low‐cost camera trap for monitoring wild insects and evaluated its performance relative to established sampling approaches. We used a low‐power microcomputer to collect time‐lapse images on colored platforms that served as non‐rewarding visual lures. We compared the performance of whole‐image and tiled deep‐learning models for automated detection of small insects in images captured by the traps. We found that tiled inference—which performs detection inside subframes of the original image—significantly improved accuracy, even outperforming human reviewers. Bumble bee visitation was highest on platforms with a repeating fluorescent bullseye pattern. Via continuous monitoring, the camera traps recorded bumble bee visits during all daylight hours over 18 days. From the recorded images, expert human taxonomists identified six Bombus species, yielding community composition and diversity estimates comparable to those obtained by hand netting. We simulated the effect of deploying variable numbers of cameras at sites with two levels of diversity. These simulations suggested that adding cameras substantially increased sampling completeness, particularly in species‐rich communities. By pairing low‐cost, semi‐automated camera traps with deep‐learning image analysis, our work provides a foundation for scalable, nonlethal studies of bumble bee diversity and activity patterns.