montimaj

montimaj / agribound

Public

An AI-powered field boundary delineation toolkit combining satellite foundation models, embeddings, and global training data for accurate agricultural parcel/field boundary mapping.

montimaj.github.ioagribound agriculture computer-vision field-boundary geoai geospatial
18
0
100% credibility
Found Mar 31, 2026 at 18 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

Agribound is an open-source Python toolkit that simplifies agricultural field boundary mapping from satellite imagery using multiple AI models and automatic land cover filtering.

How It Works

1
🗺️ Discover Agribound

You hear about a friendly tool that turns satellite photos into clear maps of farm fields, perfect for tracking crops without any hassle.

2
🔧 Set up easily

Download and prepare the tool on your computer so it's ready to map fields anywhere in the world.

3
🌍 Pick your area

Draw or select the farm region you want to study on an interactive map.

4
📅 Choose details

Select the year and satellite view that best matches your fields, plus a smart analysis method.

5
🚀 Launch mapping

Hit go, and watch as the tool pulls fresh satellite data and draws precise field outlines automatically.

6
🧹 Clean up results

The tool smooths edges, removes tiny bits, and filters to just real crop areas using land cover info.

🎉 View your fields

Enjoy beautiful maps of your delineated fields, ready to export and use for planning or reports.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 18 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is agribound?

Agribound is a Python toolkit for delineating agricultural field boundaries from satellite imagery. It pulls cloud-free composites from Landsat, Sentinel-2, or NAIP via Google Earth Engine, runs AI-powered engines like Delineate-Anything or FTW for accurate boundary detection, and outputs clean GeoPackages with automatic crop filtering using land cover data. One CLI command or API call handles the full pipeline, from global embeddings to vector polygons.

Why is it gaining traction?

It unifies seven delineation approaches—instance segmentation, ViT foundation models, unsupervised clustering—under a single reproducible workflow, skipping ad-hoc scripts. Standout: auto-removes non-agricultural polygons (roads, forests) with Dynamic World or NLCD, supports fine-tuning on local references, and scales to continents via Dask parallelism. Developers grab it for quick, no-training field maps anywhere, like Pampas soy or Indian rice paddies.

Who should use this?

GIS analysts mapping farm parcels for crop yield models, agrotech teams building water use dashboards, or remote sensing researchers tracking field changes over decades. Ideal for US NAIP high-res work, global Sentinel-2 smallholders, or embedding-based unsupervised runs without GPUs.

Verdict

Solid early alpha for agribound-style AI-powered agricultural boundary delineation—excellent docs, 15 examples across continents, Apache 2.0—but 18 stars and 1.0% credibility score mean test in sandboxes first. Worth starring if you need accurate field vectors today.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.