FarmOS

Video Demo

About this project

FarmOS is a precision agriculture platform that gives farmers AI-powered, real-time visibility into crop health, weed threats, and soil conditions — rendered inside a photorealistic 3D reconstruction of their actual farm. We went in person to real farms to capture on-site data. Using 3D Gaussian Splatting, we simulate drone flight paths through the 3D environment, extract frames, and feed them into a multi-stage Gemini pipeline: first to detect soil regions, then to identify weeds and anomalies within those regions. By combining saved 3D drone coordinates with the pixel locations Gemini flags, we place problem markers at exact real-world positions inside the 3D viewer — letting farmers (and AI agents) digitally traverse their fields before ever stepping outside. The dashboard overlays a live HUD on the 3D farm: detected weed species with treatment dispatch, pest and sprinkler alerts, ambient conditions, and a soil health summary. A dedicated Soil Dashboard visualizes 40 IoT sensors on an interactive heatmap across 5 layers (moisture, nitrogen, phosphorus, potassium, pH) using IDW interpolation. An AI report endpoint pulls live Firebase data and generates a full farm health report with per-category scores via Gemini 2.5 Flash. Three Google ADK agents handle natural language workflows: a Research Agent with live Google Search, an Operations Agent for inventory and equipment queries, and a Robot Agent for autonomous weed and spray commands. Stack: React, TypeScript, Three.js, Gaussian Splatting (WebGL), FastAPI, Firebase Firestore, Google Gemini 3.0 Flash, Google ADK.

Gallery