VISIONHIVE: An AI-Based Honeybee Entrance Monitoring and Activity Analysis System

This study presented VisionHive, a YOLOv8 and DeepSORT-based deep learning pipeline for non-invasive, automated honeybee entrance monitoring. By combining real-time detection, persistent serial-ID tracking, a virtual dual-box ROI framework and presence-based directional counting, the system addresses a gap left by prior sensor-based and detection-only approaches. Trained and validated on field data from HBRTC Pantnagar, the best model achieved a [email protected] of 0.954 and counting accuracy above 93% against manual ground truth, with noon-time lighting producing the most reliable detections. The proposed framework offers beekeepers, particularly smallholder farmers, an objective, continuous, low-cost tool to shift colony management from reactive to data-driven practice. Future work will expand the dataset across seasons and hive locations, add a dedicated pollen-carrier detection class building on prior pollen-bearing bee recognition work [14]–[15], integrate ByteTrack[3] for improved occlusion handling, deploy the model on Raspberry Pi hardware for real-time field use in line with existing IoT-based hive-monitoring platforms [12], and build a beekeeper-facing mobile interface for live hive analytics.

[1]  Bochkovskiy, A., Wang, C.Y., Liao, H.Y.M.: YOLOv4: Optimal Speed and Accuracy of Object Detection. arXiv preprint arXiv:2004.10934 (2020). [2]   Wojke, N., Bewley, A., Paulus, D.: Simple Online and Realtime Tracking with a Deep Association Metric. In: Proc. IEEE International Conference on Image Processing (ICIP), pp. 3645–3649 (2017). [3]  Nguyen, X.T., Wang, Y., Zeng, Z., Chen, S.: ByteTrack for Insect Multi-Object Tracking in Agricultural Video. In: Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pp. 3241–3249 (2022). [4]    Campbell, J., Mummert, L.B., Sukthankar, R.: Video Monitoring of Honey Bee Colonies at the Hive Entrance. In: Proc. ICPR Workshop on Visual Observation and Analysis of Animal and Insect Behavior (2008). [5]     Tu, G.J., Hansen, M.K., Kryger, P., Ahrendt, P.: Automatic Behaviour Analysis System for Honeybees Using Computer Vision. Comput. Electron. Agric. 122, 10–18 (2016). [6]  Yang, L., Liu, Y., Xu, P., Li, D.: BeeDetect: Real-Time Honeybee Detection System Using YOLOv5 for Hive Entrance Monitoring. Biosyst. Eng. 220, 1–12 (2022). [7]  Sledevič, T., Serackis, A., Matuzevičius, D., Plonis, D., Vdoviak, G.: Visual Recognition of Honeybee Behavior Patterns at the Hive Entrance. PLoS ONE 20(2), e0318401 (2025). [8]    Zhu, P., Wang, L., Zhang, C., Shi, M.: Multi-Object Tracking of Insects in Agricultural Video Using YOLO and DeepSORT. Comput. Electron. Agric. 174, 105490 (2020). [9]  Lei, C., Lu, Y., Xing, Z., Zhang, J., Li, S., Wu, W., Liu, S.: A Honey Bee In-and-Out Counting Method Based on Multiple Object Tracking Algorithm. Insects 15, 974 (2024). [10]  Simic, M., Starcevic, V., Kezić, N., Babic, Z.: Simple and Low-Cost Electronic System for Honey Bee Counting. In: Proc. 28th International Electrotechnical and Computer Science Conference, Portorož, Slovenia (2019).