HornBill

HornBill
GIF
Presenting at WRO Nationals at Hydrabad
The new Payload System with Linear Actuator System
Prototype 1 of the payload system
Flex Board
Prajesh Nair -- Drone Engineer
Aditya Tripathi -- AI Engineer
Lovely Boii <3
GIF
GIF
Nationals Golden Badge
2nd in Delhi Regionals

Inspiration

Forests are the lungs of our planet yet millions of hectares vanish every year due to deforestation, wildfires, and human activity. Manual reforestation is slow, expensive, and often impossible in remote or unsafe areas. This inspired us to create Horn-Bill, an AI-powered drone system that can autonomously plant trees by identifying suitable areas and dropping seed bombs precisely where they’re most likely to grow.

What It Does

Horn-Bill is an autonomous drone equipped with an onboard camera and AI system that identifies non-vegetated zones suitable for planting. Once a suitable area is detected, it automatically releases a biodegradable seed bomb that disperses seeds safely and effectively. The system compensates for wind direction, altitude, and drift, ensuring every seed lands exactly where it should.

How It Works (Simplified) -- refer to video

Takeoff & Stabilization: The drone takes off and enters stabilization mode to maintain altitude.
AI Vision: Using the Roboflow vegetation segmentation model, it detects green (vegetated) and brown (bare soil) regions from the live camera feed.
Wind & Physics Engine: It calculates seed drift based on wind speed, direction, and height.
Decision & Drop: If the area beneath is suitable for planting, the drone releases a seed bomb via a servo-controlled payload bay.
Simulation & Control: A Raspberry Pi handles the onboard AI, while the ground station monitors data via a real-time interface.

How We Built It

Drone Hardware: Built using lightweight carbon-fiber arms, brushless motors, and an autopilot (ArduPilot) system for both manual and autonomous flight.
AI & Vision: Implemented a vegetation segmentation model (Roboflow) integrated with OpenCV and Python for real-time inference.
Payload System: Designed a CNC-cut detachable payload box with a lever-based servo mechanism for dropping seed bombs.
Seed Bombs: Created from clay, compost, and natural fibers for weight, moisture retention, and seed protection.
Software Integration: Developed and tested the code on a Mac simulation environment, then deployed it on a Raspberry Pi connected to the drone.

Accomplishments That I'm Proud Of

🏆 2nd Place – WRO Delhi Regionals 2025
🏅 Top 10 – WRO Nationals (India)
🥇 Awarded the Golden Badge of Innovation for AI Integration

We started Horn-Bill with a vision to make reforestation faster and smarter and reaching the top ranks nationally reaffirmed our belief that technology and nature can grow together.