EN
RU
ES
ZH
FR
DE
PT
AR
HI
What is OpenCV and why is it needed?
OpenCV (Open Source Computer Vision Library) is a cross-platform open-source library designed for computer vision and machine learning tasks. It provides over 2,500 optimized algorithms that cover virtually all areas of computer vision: from simple image processing (filtering, geometric transformations) to complex tasks (face recognition, object tracking, 3D reconstruction, video analysis).
The library is written in C++ and C, which ensures high performance, and has bindings for Python, Java, MATLAB, and other languages. In this guide, we will focus on using OpenCV from C++ — this is the optimal choice for production systems where speed and efficient memory management are critical.
Why is OpenCV needed?
- Image processing: resizing, rotation, filtering, contrast enhancement, color correction.
- Video analysis: reading/writing video, motion detection, object tracking.
- Object recognition: faces, cars, text (OCR), gestures.
- Camera calibration: distortion removal, stereo vision.
- Machine learning: built-in classifiers (SVM, KNN, Decision Trees) and integration with DNN (Deep Neural Networks) for working with TensorFlow, Caffe, PyTorch models.
Installing OpenCV for C++
Unlike Python, where a single command pip install opencv-python suffices, installing OpenCV for C++ requires compilation or using pre-built binary packages. Let's look at the main methods.
1. Installation via package manager (Linux/macOS)
Ubuntu/Debian:
sudo apt updatesudo apt install libopencv-devmacOS (Homebrew):
brew install opencv2. Installation via vcpkg (Windows, Linux, macOS)
vcpkg is a cross-platform package manager from Microsoft. First, install vcpkg:
git clone https://github.com/Microsoft/vcpkg.gitcd vcpkg./bootstrap-vcpkg.sh # Linux/macOS.\\bootstrap-vcpkg.bat # WindowsThen install OpenCV:
./vcpkg install opencv4After installation, link the library in CMake:
cmake_minimum_required(VERSION 3.10)project(MyOpenCVProject)find_package(OpenCV REQUIRED)add_executable(main main.cpp)target_link_libraries(main ${OpenCV_LIBS})3. Building from source (advanced level)
This method gives maximum control over included modules and optimizations:
git clone https://github.com/opencv/opencv.gitcd opencvmkdir build && cd buildcmake -D CMAKE_BUILD_TYPE=RELEASE \ -D CMAKE_INSTALL_PREFIX=/usr/local \ -D WITH_TBB=ON \ -D WITH_V4L=ON \ -D WITH_QT=ON \ -D WITH_OPENGL=ON ..make -j$(nproc)sudo make installQuick start: loading and displaying an image
Let's verify that OpenCV is installed correctly by writing a simple program that loads an image, converts it to grayscale, and displays it.
#include <opencv2/opencv.hpp>#include <iostream>
int main() { // Load the image cv::Mat img = cv::imread("example.jpg"); if (img.empty()) { std::cerr << "Failed to load image!" << std::endl; return -1; }
// Convert to grayscale cv::Mat gray; cv::cvtColor(img, gray, cv::COLOR_BGR2GRAY);
// Display both images cv::imshow("Original", img); cv::imshow("Gray", gray);
// Wait for a key press cv::waitKey(0); cv::destroyAllWindows();
return 0;}
Also in library
Recommendations
