Building practical tools to help beekeepers detect hive issues earlier, especially in the age of Varroa
BeeSight is an early-stage prototype that monitors hive health using thermal imaging, humidity, acoustic and video sensors. The goal? Catch problems before they become disasters, without constantly opening boxes.
Hi, I'm Nimesh. I've been keeping bees since I was eight years old, when my dad and I caught a swarm in a clay pot in Ingiriya, Sri Lanka. We transferred it to a traditional hive, and I was hooked from that moment.
When Varroa arrived in Australia, everything changed. Suddenly the stakes felt higher. I found myself wishing for better tools: something that could keep watch between inspections, catch temperature drops overnight, or flag unusual patterns before they turned into bigger problems.
That's why I started building BeeSight here in Adelaide. It's an early prototype, definitely not polished or commercial-ready, but it's getting there. The current version tracks thermal patterns, humidity, acoustics, and video. I'm working on the machine learning pipeline now, along with refining the web dashboard and making sure everything actually works in a real hive.
If you're interested in following along or testing an early unit, add your email below. I'll only send updates when there's something meaningful to share.
What the current prototype does
This version combines four types of sensors to track what's happening inside the hive, without you having to open it up constantly.
Thermal imaging
Tracks brood nest temperature patterns. If the core temperature drops overnight or shifts unusually, you'll know something's up before it becomes visible.
Humidity sensing
Monitors internal humidity. Helps flag ventilation issues or changes in colony behaviour that affect moisture levels.
Acoustic monitoring
Listens to hive sounds. Early testing shows it can pick up on changes in buzz patterns that might indicate stress or queenlessness.
Video capture
Records bee traffic at the entrance. Useful for estimating colony strength and catching unusual activity without opening the hive.
Real things BeeSight can detect
Brood temperature drops — Overnight cooling in the brood nest that might indicate weak colony strength or failing queen
Traffic pattern changes — Sudden drop in bee traffic that could signal swarming, disease, or other issues
Unusual acoustic signatures — Changes in hive sound that experienced beekeepers recognise but are hard to monitor 24/7
Humidity spikes — Could indicate ventilation problems, moisture buildup, or failing hive structure
Dataset for Varroa detection — Right now I'm gathering thermal and video data to eventually train models that can spot mite patterns
Note: This is an early prototype. Not all features are reliable yet, and I'm still iterating on what actually works in the field versus what sounded good on paper.
Where things are at
Current status of the project—what's working, what I'm testing, what's coming next.
✓Built and working
- •Prototype v0.3 printed and assembled
- •Thermal camera running, capturing frames every 10 minutes
- •Humidity and temperature sensors integrated
- •Web dashboard shows live data and basic charts
- •Over-the-air updates working in internal tests
- •Video pipeline capturing entrance activity
→In progress
- •Refining acoustic monitoring (lots of noise to filter out)
- •Building ML pipeline for thermal anomaly detection
- •Testing battery life (targeting 2-3 months between charges)
- •Early field test planned for January with three hives
- •Multi-hive dashboard view (so you can compare across hives)
Want updates as I build this?
I'll email you when there's real progress to share: field test results, new prototype versions, or when I'm ready for pilot testers. No marketing fluff, just honest updates from the workshop.
Building BeeSight in Adelaide, South Australia
If you're a researcher, beekeeper, or just curious about the project, feel free to reach out. Always happy to chat about bees, sensors, or the challenges of keeping Varroa at bay.