TerraGrow
Geo-Spatial Farmland Intelligence
● frontend-only mode

Upload satellite imagery.
Get Farm / Barren / City / Water breakdown.

TerraGrow segments every pixel into 4 classes and returns percentage coverage. Works 100% in-browser. Optionally connect the Python backend for batch + your trained model.

Aggregate coverage

Farmland —
Barren —
City / Urban —
Water —
0 images · 0 px

Drag & drop satellite / drone images here

or click Upload. Supports PNG / JPG / TIFF-converted. Multiple files = averaged + per-image table.

Farm = greens Barren = browns / tan City = gray / white / rooftops Water = blue / teal

Farm dashboard · Cloudflare Worker API (Phase 2/4)

● cloud not tested
1. Register plot
2. Advise & monitor
3. Cloud history

No cloud data loaded yet.

Connect a Worker (Test ↑) then create/select a farm.

How analysis works

1. Upload

Images stay in your browser. Canvas downsamples to max 512px for speed.

2. Segment

Each pixel → Farmland / Barren / City / Water by calibrated color rules that mimic your trained CV model. Swap in your real model in app.py → MODEL_PATH.

3. Report

Percentages, masked overlay, confidence note, CSV export for GIS records.

To use your trained model: place model.onnx / model.h5 next to app.py and set USE_TRAINED_MODEL = True. API returns identical JSON so the frontend needs no changes.