This repository provides building blocks for developing custom Python applications for Photoneo 3D sensors. Two independent integration paths are available depending on which interface you want to use.
Python examples using the phoxi-api package — a high-level Python wrapper for the PhoXi Control application. This is the recommended starting point for most use cases.
What you will find:
- Listing available devices
- Freerun and software-triggered frame acquisition
- Interactive point cloud visualisation with Open3D
- Reading and writing device settings via attribute-style and path-based APIs
- Reading offline scan files (
.praw/.pmraw): metadata, matrices, texture normalisation, and two-file comparison - MaintenanceTool API for laser power calibration and correction patching, including a robot-controlled variant (TCP server/client, with a Universal Robots example client)
Requirements: PhoXi Control >= 1.17.0 must be installed and running. Uses uv as the package manager.
→ PhoXi-API examples and quick start
Python examples using the GigE Vision / GenICam protocol directly via the harvesters library and a GenTL producer. This path gives low-level access to the device without PhoXi Control.
Requirements: Balluf ImpactAcquire GenTL producer and the GENICAM_GENTL64_PATH environment variable. Firmware >= 1.13.0.
Straightforward scripts with no visualisation dependencies. Good starting point for understanding the GenICam feature and component model.
What you will find:
- Device discovery and status inspection (components, chunks, trigger mode)
- Single-frame grab with software trigger
- Continuous freerun acquisition with FPS measurement
- Reading and writing GenICam features including raw memory registers
- Reading chunk data (temperature, calibration matrix, distortion coefficients)
- User set save/load
- Network diagnostics and Jumbo Frames configuration
More elaborate examples using Open3D and OpenCV. Include a shared photoneo_genicam helper package for components, chunks, point cloud assembly, and visualisation.
What you will find:
- Point cloud capture, visualisation, and PLY export
- Point cloud with normals and texture colour mapping
- Coordinate space transformation into marker space via chunk data
- Real-time point cloud streaming in ProjectedC mode (client-side XYZ reconstruction from depth maps)
- Coloured point cloud using color camera texture
- Confidence map and depth map display
- All texture sources captured and saved for comparison
- Hardware trigger and hardware-synchronised multi-device PTP timestamps
- Color settings ROI mode
- YCoCg color decoding (Numba JIT)
- User set management
- Marker Dot Correction workflow (feature-accessor and raw-register approaches)
- Feature effect demonstration (ISO, HDR, ProjectionOffset)
A standalone, dependency-free Python implementation of the GigE Vision discovery and FORCEIP commands for listing devices on the network and temporarily changing their IP configuration.
Visit www.photoneo.com for the most up-to-date documentation. For issues with these examples contact the Photoneo Support team at the Help Center.
Photoneo examples are distributed under the BSD License.