OpenCV
The computer vision library — image processing, detection, tracking and calibration.
What it is
OpenCV (Open Source Computer Vision Library) is a powerful C++ library for computer vision, image processing, and machine learning. It provides a wide range of algorithms for real-time image and video analysis.
OpenCV in C++ provides classes and functions to read, write, process images and videos, and perform advanced computer vision tasks. It supports matrices (cv::Mat), image filters, feature detection, machine learning, and GPU acceleration.
- Licence
- Apache 2.0
- Watch for
- OpenCV reads images as BGR, not RGB — a perennial source of confusion
When to use it
The question documentation cannot answer for you — because it cannot recommend something else.
Reach for it when
- Any image or video processing task: filtering, feature detection, camera calibration, tracking
- You need real-time performance with optional GPU acceleration
Look elsewhere when
- You only need to load and resize images — a smaller library will do without OpenCV's size
- The task is purely deep-learning inference, where a dedicated runtime may fit better
Installation
sudo apt install libopencv-devGetting started
The smallest useful thing you can do with it, and what each part means.
#include <opencv2/opencv.hpp>
int main() {
cv::Mat img = cv::imread("image.jpg");
if(img.empty()) return -1;
cv::imshow("Image", img);
cv::waitKey(0);
return 0;
}#include <opencv2/opencv.hpp>
int main() {
cv::Mat img = cv::imread("image.jpg");
cv::Mat gray;
cv::cvtColor(img, gray, cv::COLOR_BGR2GRAY);
cv::imshow("Gray", gray);
cv::waitKey(0);
return 0;
}Advanced usage
Where the library earns its place over a simpler alternative.
#include <opencv2/opencv.hpp>
int main() {
cv::Mat img = cv::imread("image.jpg", cv::IMREAD_GRAYSCALE);
cv::Mat edges;
cv::Canny(img, edges, 100, 200);
cv::imshow("Edges", edges);
cv::waitKey(0);
return 0;
}#include <opencv2/opencv.hpp>
int main() {
cv::VideoCapture cap(0);
if(!cap.isOpened()) return -1;
cv::Mat frame;
while(true) {
cap >> frame;
if(frame.empty()) break;
cv::imshow("Webcam", frame);
if(cv::waitKey(1) == 'q') break;
}
return 0;
}#include <opencv2/opencv.hpp>
int main() {
cv::Mat img = cv::Mat::zeros(512,512,CV_8UC3);
cv::line(img, cv::Point(0,0), cv::Point(511,511), cv::Scalar(255,0,0), 5);
cv::rectangle(img, cv::Point(100,100), cv::Point(300,300), cv::Scalar(0,255,0), 3);
cv::circle(img, cv::Point(256,256), 50, cv::Scalar(0,0,255), -1);
cv::imshow("Shapes", img);
cv::waitKey(0);
return 0;
}#include <opencv2/opencv.hpp>
int main() {
cv::CascadeClassifier face_cascade;
face_cascade.load("haarcascade_frontalface_default.xml");
cv::Mat img = cv::imread("group.jpg");
cv::Mat gray;
cv::cvtColor(img, gray, cv::COLOR_BGR2GRAY);
std::vector<cv::Rect> faces;
face_cascade.detectMultiScale(gray, faces);
for(auto &f : faces) cv::rectangle(img, f, cv::Scalar(255,0,0), 2);
cv::imshow("Faces", img);
cv::waitKey(0);
return 0;
}Errors and fixes
The failures you are most likely to hit, and what actually resolves them.
- cv::Exception: OpenCV(…): error
- Check image paths and ensure the file exists before reading.
- Segmentation fault
- Ensure Mat objects are initialized and not empty before performing operations.
- CascadeClassifier load failed
- Verify the path to Haar cascade XML files and that they are accessible.
Best practices
- Use cv::Mat for all image operations for performance and flexibility.
- Release resources properly and close windows to prevent memory leaks.
- Use proper color space conversions when needed (BGR, RGB, GRAY).
- Optimize pipelines for real-time applications using GPU or multi-threading.
- Keep pre-trained models and classifiers organized for reuse.
Alternatives
Comparable options, and the reason you would pick one over the other.
Background
Why it exists, and what it was reacting to.
OpenCV was originally developed by Intel in 1999 to advance computer vision research. It provides highly optimized algorithms for image processing, object detection, facial recognition, and more. OpenCV is widely used in industry and research for real-time applications and robotics.