imgproc

Imgproc is a module of the OpenCV library that provides a comprehensive set of image processing and computer vision algorithms. It offers functions for image filtering, color space conversions, geometric transformations, morphology, edge and contour analysis, and feature detection. The module is designed to be efficient and portable, functioning on dense image arrays in C++ (Mat) and on NumPy arrays in Python (via the cv2 interface).

Key areas covered by Imgproc include spatial filtering (such as blur, GaussianBlur, medianBlur, and bilateralFilter), sharpening and smoothing operations, and a full set of morphological operations (erode, dilate, opening, closing, and related transforms). It also includes gradient and edge detectors (Sobel, Scharr, Laplacian, and Canny), as well as color space conversions (for example, BGR to gray or HSV) and thresholding (threshold, adaptiveThreshold, inRange).

Imgproc provides geometric image transformations, including resizing, affine and perspective warping (warpAffine, warpPerspective) and remapping, which are essential for data augmentation and alignment tasks. It also supports histogram operations and image equalization.

Contour and shape analysis form a core part of Imgproc, with functions for finding and drawing contours (findContours, drawContours), computing moments and areas (contourArea, arcLength), and shape fitting (approxPolyDP, fitEllipse, fitLine). The module also includes Hough transform-based detectors for lines and circles (HoughLines, HoughCircles), enabling basic geometric feature extraction.

Imgproc is accessible across language bindings in OpenCV, with its C++ declarations in opencv2/imgproc.hpp and Python exposure via cv2. It serves as a foundation for preprocessing, segmentation, feature extraction, and various vision pipelines.