Machine Vision Toolbox for Python
Object-oriented vision for Python. A unified framework for spatial reasoning and computer vision harmonizing NumPy, OpenCV, and Open3D.
Explore the rationale, see code examples, and get the toolbox installed in your local environment
Access over 100 specialized functions for image filtering, feature extraction (blobs, points, lines), mathematical morphology, stereo vision, bundle adjustment, camera calibration and more.
The object-oriented heart of the toolbox. High-level wrappers for Image, PointCloud, and Camera classes, built on a robust computational base.
Learn by doing with our JupyterLite environment. Run examples instantly in your browser or explore our library of Google Colab templates.
Bridge the gap to the wider robotics ecosystem with native support for ROS 2 message streams and PyTorch tensor interfaces.
Pythonic OpenCV Wrapper: Images are first-class objects with intuitive operators and transparent BGR/float32 handling.
100+ Vision Tools: From image acquisition and morphology to blob, point, and line feature extraction.
Geometric Intelligence: Purpose-built for vision-based control, including visual Jacobians, homographies, and camera calibration.
Efficient Core: Inherits the performance and maturity of the NumPy and OpenCV ecosystems.