Inspiration
We were inspired by our experiences struggling to care for plants, where we'd find ourselves at a loss for why our plants died or fell ill. Then we thought, what if it could just tell us what it needed.
What it does
Our G.R.O.O.T. (Grow, Respond, Observe, Optimize, Tend) system enables human plant interaction with the following features:
- Voice : Talk with your plant to get actionable care advice ("I'm thirsty!").
- Personality : Tailor its personality, from peaceful pothead to sassy succulent.
- Sensing Suite : Detects key environmental metrics to give you best advice on how to care for your plant, from soil moisture to blemish detection via computer vision.
- Face Display : A digital display on the pot reflects your plant's voice and emotions.
- Vacation automation : Automated watering for while you're away from home.
How we built it
G.R.O.O.T. utilizes a 3-layer architecture consisting of:
- Hardware -> Viam Server (MCU + MPU)
- Middleware -> Viam Cloud Service
- Intelligence -> Agentic & Generative AI + Computer Vision
Challenges we ran into
The biggest challenges we ran into involved limitations in hardware and network capabilities.
- Hardware -> Limited resolution in available sensors & ADC (e.g., soil moisture)
- Network -> Limited internet connectivity (e.g., Raspberry Pi Reliability Issues)
Accomplishments that we're proud of
- Voice -> Proud of the personality we were able to craft.
- Computer Vision -> Proud of the information we were able to procure.
- Team Spirit -> The team brought constant energy throughout the entirety of the hackathon, even during setbacks.
What we learned
- Computer Vision Modelling
- Geographical Information Systems
- Hardware-Software Interfacing
What's next for Leaf It To Us
- LLM Reasoning Improvements
- Hardware Failsafes (e.g., loss of Wi-fi connectivity)
- Full ecosystem: plants and autonomous robots in harmony.
Log in or sign up for Devpost to join the conversation.