# trailbehind/DeepOSM

Train a deep learning net with OpenStreetMap features and satellite imagery.

Repository: https://github.com/trailbehind/DeepOSM
Canonical: https://ross.abutalabs.com/products/deeposm
Language: Python
License: MIT
License Family: permissive
Last push: 2017-03-13T16:09:05+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3853, "days_push": 3460, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1330, forks 184 (observed 2026-08-28T04:04:24.226329+00:00)

## What it is
DeepOSM is a Python application that trains neural networks with TensorFlow to classify roads and features in satellite imagery using OpenStreetMap data as training labels. It downloads NAIP satellite imagery and OSM PBF extracts, generates training data, and renders predictions of mis-registered roads as JPEGs.

## Use cases
- train a neural network to detect roads in satellite imagery
- find errors in OpenStreetMap road data
- generate training data from OSM and NAIP imagery
- classify features in aerial photos with deep learning
- compare OSM data against satellite imagery predictions

## When to choose
- you want to train a deep learning model on OpenStreetMap labels and satellite imagery
- you need to identify mis-registered or missing roads in OSM data
- you want a ready-made TensorFlow pipeline for geospatial imagery classification

## When to avoid
- you need actively maintained software or recent TensorFlow compatibility
- you want general-purpose satellite image analysis beyond road/feature classification
- you cannot provide AWS credentials to download requester-pays NAIP imagery

## Facets
- artifact type: application
- maturity: abandoned
- function: machine-learning, deep-learning, computer-vision, image-processing, geospatial
- domain: machine-learning, computer-vision
- platform: python
- tags: openstreetmap, satellite-imagery, tensorflow, naip, road-detection, neural-networks, maps, geospatial, linux, macos, docker

## Member repositories
- trailbehind/DeepOSM (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:24.226329+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T04:45:37.036800+00:00, confidence not recorded.
  - readme: https://github.com/trailbehind/DeepOSM (fetched 2026-08-28T04:04:24.226329+00:00, sha ccd1383504dd)
- Data as of 2026-08-30T08:39:29.467469+00:00.