Research library

Physics-aware vision instrumentation for stingless bee counting at hive entrance using hybrid edge-cloud object detection

Stingless bee entrance monitoring requires a non-invasive tool to measure colony traffic without disrupting foraging. This paper presents a hybrid edge-cloud object detection and physics-aware vision framework using a Raspberry Pi 4 edge node, Sony IMX296 global-shutter camera, YOLO11 detection, and OC-SORT tracking. The analysis shows how frame rate, bee velocity, field-of-view scale, detector recall, and tracker association tolerance jointly affect counting reliability. Validation on 14 one-minute Geniotrigona thoracica entrance videos achieved 85.0% IN accuracy, 66.7% OUT accuracy, and 76.3% total counting accuracy against manual counts.

Download PDF

Publication details

Authors
Mohd Amri Md Yunus, Lari Andres Sanjaya, Celestine Hiu Shun Yi, Shafishuhaza Sahlan, Saharudin Saharudin, Sudrajat Harris, Hendri Maja Saputra
Organizations
🇲🇾 Universiti Teknologi Malaysia🇮🇩 Universitas Negeri Jakarta🇮🇩 Institut Teknologi Indonesia🇮🇩 Politeknik Negeri Bandung🇮🇩 National Research and Innovation Agency
Year
2026
Type
Conference

Relevancy to Gratheon

Although evaluated on stingless bees, the instrumentation and tracking constraints transfer directly to Gratheon's honey-bee gate tracker. The paper quantifies a failure mode that entrance counters often overlook: at 15 FPS, fast ingress can move a bee beyond the tracker's association distance between frames, fragmenting tracks and biasing directional counts. Its global-shutter camera choice, edge-cloud split, and explicit relationship between velocity, frame rate, field of view, and OC-SORT tolerance provide concrete guidance for camera selection, edge inference budgets, and field validation of bee traffic analytics.