Project Details

[Return to Previous Page]

LumiFlor: an adaptive, autonomous, control system for real-time bee detection and monitoring

Company: PSU Department of Entomology - Grozinger Lab

Major(s):
Primary: EE
Secondary: CMPEN
Optional: CMPSC, ME

Non-Disclosure Agreement: NO

Intellectual Property: NO

Bees are essential for global agriculture and ecosystems, yet traditional monitoring relies on slow, labor-intensive manual sampling and observation. To address this, we are engineering LumiFlor, an autonomous Cyber-Physical System designed to actively track, behaviorally evaluate, and identify bee species in real time. 1. Active LED Attraction: Transitioning from passive traps to active monitoring, LumiFlor utilizes programmable, custom-built LED arrays tailored to bee visual biology, emitting specific wavelengths to safely attract them. 2. Array Design: Fabricated entirely in-house using a desktop PCB printer, four interchangeable artificial "flowers" dynamically switch sizes and precise UV, blue, and green wavelength combinations to support rapid behavioral studies. 3. Camera Monitoring System: A stereoscopic pair of high-speed cameras (over 500 fps) quantitatively tracks complex 3D foraging behaviors. Simultaneously, it extracts synchronized image bursts for automated, real-time species classification. 4. Power Systems and Autonomous Operation: To ensure long-term field autonomy, a solar-harvesting suite employs "graceful degradation," dynamically cutting power to non-essential peripherals during low-energy states to maintain core sensing functions. 5. Edge Computing Integration: Adapting our existing YOLO-based AI pipeline, the onboard system processes rapid motion detection and object tracking, instantly modifying the data output for deep scientific analysis. By merging dynamic floral mimicry with edge AI, LumiFlor delivers an unprecedented, labor-reducing hardware platform for real-time conservation and agricultural research. 1. LumiFlor Physical Prototype: A fully assembled, weather-resistant field unit that combines the calibrated stereoscopic high-speed camera rig with the programmable LED arrays. 2. LED Arrays, Files, & Firmware: Fully fabricated UV/blue/green PCB boards, accompanied by the control firmware for floral patterns and printable hardware design files (ECAD/Gerber). 3. Protocols & Datasets: Standardized experimental assay protocols, a field validation dataset, and an ethological database linking species classifications with complex 3D behavioral metrics. 4. Power Management Unit (PMU): An integrated solar harvesting and battery system. 5. Edge-Computing & AI Pipeline: A ruggedized hardware module running the integrated classification software to extract synchronized images and identify species in real time.