ORBIT - OpenDrive Road Builder from Imagery Tool

A visual tool for creating or editing ASAM OpenDRIVE 1.8 road networks from aerial imagery.

Note

This open source project is maintained by RISE Research Institutes of Sweden. See LICENSE file for open source license information.

Note

This is a beta version. Bugs and missing features should be expected. Github issues can be added for bug reports or feature requests.

---

Contents

• Features
• Installation
• Quick Start
• Documentation
• Related Packages
• Project Structure
• Development
• License

---

Features

Road Annotation

• Interactive polyline drawing on aerial/satellite/drone images
• Centerline and lane boundary distinction with road mark types (solid, broken, etc.)
• Lane sections for roads where lane configuration changes, with split and merge operations
• Road splitting and merging for flexible network editing
• Data-driven road marks from actual annotated line types
• OpenDRIVE 1.8 lane attributes (direction, advisory)

Junction Support

• Junction annotation with drag-and-drop positioning
• Roundabout wizard for creating circular intersections
• Connecting roads with proper geometric paths through junctions
• Lane-level connections with explicit lane-to-lane mappings
• Automatic connection generation from road geometry

Import Capabilities

• OpenStreetMap import via Overpass API (roads, signals, junctions, objects, land use areas)
• OpenDRIVE import for editing existing .xodr files (round-trip support)

Georeferencing

• Control point system for pixel-to-geographic transformation
• CSV import for batch control points
• Monte Carlo uncertainty analysis with visualization
• Validation metrics with reprojection error

Export

• ASAM OpenDRIVE 1.8 XML format
• XSD schema validation against official ASAM schema (download)
• Configurable geometry — preserve all points or fit curves
• Geographic reference with PROJ4 projection string
• Complete junction export with connecting roads and lane links

---

Installation

Using uv (recommended)

# Install uv if needed
curl -LsSf https://astral.sh/uv/install.sh | sh

# Clone and install
git clone https://github.com/RI-SE/ORBIT.git
cd ORBIT
uv sync

Using pip

python3 -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate
pip install -e .

---

Quick Start

# Start with an image
orbit path/to/aerial_image.jpg

# Start empty (load image via File menu)
orbit

# Enable verbose logging
orbit --verbose

# Enable XSD schema validation for exports
orbit --xodr_schema /path/to/OpenDRIVE_Core.xsd

Note: After installation with uv sync or pip install -e ., the orbit command is available directly. Alternatively, use uv run orbit or python run_orbit.py.

Basic Workflow

1. Load image — File → Load Image or pass path on command line
2. Add control points — Tools → Georeferencing (minimum 4 points (oblique imagery) or 3 (nadir imagery))
3. Import or draw — Either import an existing map from OpenStreetMap or OpenDRIVE, or draw roads directly in ORBIT.
4. Edit — Edit roads. Add signs, parking, land use areas, and objects.
5. Export — File → Export → Export to OpenDrive

---

Documentation

Guide	Description
User Guide	Complete user guide with workflow, tips, and keyboard shortcuts
Georeferencing Guide	Control points and uncertainty analysis
OSM Import Guide	OpenStreetMap import feature
Validation Guide	Validation metrics and uncertainty estimation
Developer Guide	Architecture and contribution guidelines

---

Related Packages

All three are MIT licensed and free of any PyQt dependency, so they can be used in scripts, pipelines and downstream tools without taking on ORBIT's GPL-3.0 terms.

orbit-core

The headless half of ORBIT: road-network model, OpenStreetMap and OpenDRIVE import, and OpenDRIVE export. Build a CARLA-ready .xodr from coordinates in three calls, with no GUI and no imagery:

import orbit_core as oc

bbox = oc.bbox_from_center(57.6968, 11.9865, radius_m=150)
osm = oc.osm_data_from_overpass(bbox)      # or osm_data_from_file("area.osm")
oc.opendrive_from_osm_data(osm, bbox, "out.xodr")

See orbit-core/README.md for details.

opendrive-map

Read-only OpenDRIVE road-network and lane-geometry model — parses a .xodr into lane polygons, centrelines and junctions, with point-to-lane lookup. Shared across the toolchain (COSMO's trajectory-explorer, data-metrics). See opendrive-map/README.md.

orbit-georef

Standalone Python library for pixel↔geo coordinate transformation. Use it to work with ORBIT's georeferencing outside the GUI — for example, converting pixel coordinates to lat/lon in scripts or downstream tooling. See orbit-georef/README.md for details.

---

Project Structure

orbit/
├── gui/          # PyQt6 GUI (MainWindow, ImageView, dialogs, widgets)
└── signs/        # Traffic sign libraries (country-specific)
orbit-core/       # Headless core, MIT (separate package)
└── src/orbit_core/
    ├── models/       # Data models (Road, Polyline, Junction, ParkingSpace, Signal, ...)
    ├── importers/    # OSM and OpenDRIVE importers
    ├── export/       # OpenDRIVE XML generation (writers, builders)
    └── utils/        # Coordinate transforms, geometry utilities
opendrive-map/    # Read-only OpenDRIVE lane geometry, MIT (separate package)
orbit-georef/     # Standalone georeferencing library, MIT (separate package)

The GUI is a shell over orbit-core: all road-building logic lives in the library, and nothing in it imports PyQt.

Project Files

Projects save as .orbit JSON files containing:

• Image path and metadata
• Polylines (pixel coordinates)
• Roads with lane sections
• Junctions with connections and junction groups
• Control points for georeferencing
• Signals and roadside objects
• Parking spaces
• Land use areas (forest, farmland, water, etc.)

---

Development

Setup

# Install with dev dependencies
uv sync --extra dev

# Run tests
uv run python -m pytest tests/ -v

Key Technologies

• PyQt6 — GUI framework
• NumPy/SciPy — Geometry and transformations
• lxml — XML generation
• pyproj — Coordinate projections
• xmlschema — OpenDRIVE XSD validation

See Developer Guide for architecture details.

---

License

The main ORBIT project is licensed under the GNU General Public License v3.0 (GPL-3.0).

The separate libraries orbit-georef (located in orbit-georef/) and opendrive-map (located in opendrive-map/) are licensed under the MIT License — see orbit-georef/LICENSE and opendrive-map/LICENSE — allowing for more permissive use in downstream projects. Neither library depends on PyQt6, which is what places the main application under GPL-3.0.

Dependencies and Their Licenses

Main ORBIT project (runtime):

• PyQt6 - GPL v3 (commercial license available)
• PyQt6-Qt6 - LGPL v3 (Qt framework bindings)
• opencv-python - Apache 2.0
• NumPy - BSD 3-Clause License
• SciPy - BSD 3-Clause License
• lxml - BSD 3-Clause License
• pyproj - MIT License
• xmlschema - MIT License

Main ORBIT project (development, optional):

• pytest - MIT License
• pytest-cov - MIT License
• pytest-mock - MIT License
• ruff - MIT License

orbit-georef library (runtime):

• NumPy - BSD 3-Clause License
• pyproj - MIT License

orbit-georef library (development, optional):

• pytest - MIT License
• pytest-cov - MIT License

opendrive-map library (runtime):

• NumPy - BSD 3-Clause License
• Shapely - BSD 3-Clause License
• pyproj - MIT License

opendrive-map library (development, optional):

• pytest - MIT License

Acknowledgement

This package is developed as part of the SYNERGIES project.

Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or European Climate, Infrastructure and Environment Executive Agency (CINEA). Neither the European Union nor the granting authority can be held responsible for them.