FlyCount: high-speed counting of black soldier flies using neuromorphic sensors

Black Soldier Fly larvae (BSFL) play a vital role in waste management by efficiently converting organic waste into protein-rich feed, thereby reducing environmental impact and supporting sustainable agriculture. To optimize BSFL production, precise monitoring of adult flies is essential. This paper introduces FlyCount, a system developed with neuromorphic vision sensors and a custom spike detection algorithm for real-time, accurate fly counting. The system’s architecture integrates advanced event stream processing with dynamic thresholding, achieving 95% accuracy across 30 trials, validated against manual counts. FlyCount also visualizes fly trails, providing real-time insights into movement patterns. Moreover, variations in count plots enable monitoring of peak fly departures from the hatch, offering critical data for optimizing hatching and collection processes. This data-driven approach enhances the scalability of waste processing systems, fostering sustainability and a brighter future, while also representing a foundational step toward a broader goal of leveraging neuromorphic systems to tackle the complex challenges facing humanity.