OpenCV

OpenCV (Open Source Computer Vision Library) is the world's most widely used computer vision and machine learning software library, originally developed by Intel in 1999 by Gary Bradski and maintained by Itseez (acquired by Intel in 2016). With over 79,000 stars on its primary repository as of 2026, OpenCV provides over 2,500 optimized algorithms for real-time computer vision across C++, Python, Java, and MATLAB interfaces. Core modules include: imgproc (image processing: filtering, geometric transformations, color space conversions, histograms, thresholding, contour detection, morphological operations), imgcodecs (image I/O for JPEG, PNG, TIFF, WebP, HDR, EXR), videoio (video capture and writing), highgui (GUI windows, trackbars, mouse events), dnn (deep learning inference with support for ONNX, TensorFlow, PyTorch/Torch, Caffe, Darknet models, and CUDA backend acceleration), features2d (feature detection and description: SIFT, SURF, ORB, AKAZE, BRISK, FAST, BRIEF), calib3d (camera calibration, stereo vision, pose estimation, epipolar geometry, structure from motion), objdetect (object detection: Haar cascades for faces and eyes, HOG descriptor for pedestrians, QR code and ArUco marker detection), video (motion estimation, optical flow: Farneback, Lucas-Kanade, DeepFlow, background subtraction: MOG2, KNN), photo (computational photography: HDR, seamless cloning, non-photorealistic rendering, inpainting, despeckling), stitching (image panorama creation), ml (traditional ML: SVM, kNN, Decision Trees, Random Forest, Boosting, ANN MLP), and cuda modules for GPU acceleration. OpenCV runs on Linux, Windows, Android, iOS, and macOS.