OpenCV (Open Source Computer Vision Library)

The most widely used open-source computer vision library, for image, camera, and geometry processing.

OpenCV is an open-source computer vision library originally started by Intel, with its core written in C++ and bindings for Python and other languages; in Python it's used via import cv2. It covers image reading and resizing, color-space conversion, filtering, feature-point extraction and matching, camera calibration, distortion correction, PnP pose estimation, and ArUco marker detection, among other classic vision tasks. Deep learning has taken over most recognition tasks, but OpenCV still gets used constantly in robotics: reading camera frames, preprocessing, calibrating camera intrinsics and extrinsics, and visualizing results — it's hard to avoid.

ExampleUse cv2.calibrateCamera with a checkerboard pattern to calibrate a wrist camera's intrinsics, then cv2.undistort to remove lens distortion.

Also called

cv2

Related

Camera Calibration · ArUco Marker · Perspective-n-Point · Feature Points · Point Cloud Library (PCL) · Open3D

Sources

OpenCV - Open Computer Vision Library