📊 Easily Plot with CVPlot: The “Mini Matplotlib” in C++

Are you using C++ for image processing, robotics, scientific computing, or algorithm experimentation, but struggling with:

“I want to draw a line chart, scatter plot, or display data in real-time, but C++ doesn’t have a plotting tool as simple and easy to use as Python’s Matplotlib?”

Don’t worry! Today we will introduce an lightweight, easy-to-use, and GUI framework-independent open-source C++ plotting library — CVPlot.

🔍 What is CVPlot?

CVPlot is an open-source C++ plotting library designed specifically for computer vision and numerical experiments. Its name comes from “CV” (Computer Vision) + “Plot” (plotting), but its applications extend far beyond CV.

✅ Core Features:

Feature	Description
🎯 Lightweight and Simple	Header-only, no compilation required
🖼️ Depends on OpenCV	Uses cv::Mat to display images, seamless integration
📈 Supports Various Charts	Line charts, scatter plots, bar charts, heatmaps, etc.
🔄 Real-time Plotting	Supports dynamic updates, suitable for visualizing algorithm processes
📦 No External GUI Dependencies	Directly displays using OpenCV’s imshow

Project address: https://github.com/pkuros/cvplot

🛠️ Installation and Configuration (Ubuntu/Linux Example)

CVPlot is a header-only library, which means you only need to download the header file to use it.

1. Install OpenCV (Prerequisite)

sudo apt-get install libopencv-dev

2. Download CVPlot Header File

git clone https://github.com/pkuros/cvplot.git

You will get a cvplot.h file, just place it in your project directory.

💡 First Example: Plotting a Sine Wave

Let’s use CVPlot to draw a simple sine function graph.

✅ Project Structure

cvplot_demo/
├── cvplot.h
├── main.cpp

✅ main.cpp Code

#include <iostream>
#include <vector>
#include <cmath>
#include "cvplot.h"  // Include CVPlot
#include <opencv2/opencv.hpp>

int main() {
    // 1. Prepare data: x from 0 to 4π, y = sin(x)
    std::vector<float> x, y;
    for (float t = 0; t <= 4 * M_PI; t += 0.1) {
        x.push_back(t);
        y.push_back(std::sin(t));
    }

    // 2. Use CVPlot to plot
    cvplot::Figure fig;
    fig.plot(x, y, cvplot::PlotStyle::Line, cv::Scalar(255, 0, 0)); // Red line
    fig.title("y = sin(x)");
    fig.xlabel("x");
    fig.ylabel("y");

    // 3. Display image (using OpenCV)
    cv::Mat plot = fig.render(600, 400);  // Width 600, height 400
    cv::imshow("Sine Wave", plot);
    cv::waitKey(0);  // Wait for key press to close

    return 0;
}

✅ Compilation Command

g++ main.cpp -o sine_plot `pkg-config --cflags --libs opencv4`  # OpenCV 4
# or
g++ main.cpp -o sine_plot `pkg-config --cflags --libs opencv`   # OpenCV 2/3

🖨️ Result

You will see a window displaying a red sine wave with a title and axis labels.

🧮 Going Further: Plotting Scatter Plot + Multiple Curves

Let’s plot two sets of data: sine and cosine, and display them using a scatter plot.

#include <vector>
#include <cmath>
#include "cvplot.h"
#include <opencv2/opencv.hpp>

int main() {
    std::vector<float> x, y1, y2;
    for (float t = 0; t <= 2 * M_PI; t += 0.2) {
        x.push_back(t);
        y1.push_back(std::sin(t));
        y2.push_back(std::cos(t));
    }

    cvplot::Figure fig;

    // Plot sin(x) as a blue line
    fig.plot(x, y1, cvplot::PlotStyle::Line, cv::Scalar(255, 0, 0));

    // Plot cos(x) as green points
    fig.plot(x, y2, cvplot::PlotStyle::Points, cv::Scalar(0, 255, 0));

    fig.title("Sine vs Cosine");
    fig.xlabel("x (radians)");
    fig.ylabel("y");
    fig.legend({"sin(x)", "cos(x)"});

    cv::Mat plot = fig.render(700, 500);
    cv::imshow("Trig Functions", plot);
    cv::waitKey(0);

    return 0;
}

🖨️ Result

• Blue line: sin(x)
• Green points: cos(x)
• With legend, title, and axes

Very suitable for comparing algorithm outputs or data distributions!

📊 Supported Chart Types

CVPlot supports various plotting methods:

Chart Type	Method
Line Chart	fig.plot(x, y, PlotStyle::Line)
Scatter Plot	fig.plot(x, y, PlotStyle::Points)
Bar Chart	fig.bar(labels, values)
Heatmap	fig.imshow(matrix) (requires cv::Mat)
Text Annotation	fig.text(x, y, "label")

Example: Bar Chart

std::vector<std::string> labels = {"A", "B", "C", "D"};
std::vector<float> values = {3.5, 2.0, 4.8, 1.2};

cvplot::Figure fig;
fig.bar(labels, values, cv::Scalar(0, 128, 255));
fig.title("Bar Chart Example");
fig.ylabel("Score");

cv::Mat bar_plot = fig.render(500, 400);
cv::imshow("Bar Chart", bar_plot);
cv::waitKey(0);

⚙️ Advanced Techniques: Real-time Dynamic Plotting

CVPlot is also suitable for real-time visualization, such as sensor data streams or algorithm iteration processes.

cvplot::Figure fig;
cv::namedWindow("Real-time Plot", cv::WINDOW_AUTOSIZE);

for (int frame = 0; frame < 100; ++frame) {
    std::vector<float> x, y;
    float phase = frame * 0.1;
    for (float t = 0; t < 4 * M_PI; t += 0.2) {
        x.push_back(t);
        y.push_back(std::sin(t + phase));  // Dynamic phase
    }

    fig.clear();
    fig.plot(x, y, cvplot::PlotStyle::Line, cv::Scalar(0, 255, 255));
    fig.title("Real-time Sine Wave");

    cv::Mat plot = fig.render(600, 400);
    cv::imshow("Real-time Plot", plot);
    cv::waitKey(100);  // Pause 100ms per frame
}

You will see the sine wave “moving”! Very suitable for debugging filters, control algorithms, etc.

✅ Summary of Advantages

Advantage	Description
📦 Lightweight	Only one header file, no compilation dependencies
🧩 Easy Integration	Based on OpenCV, suitable for CV/robotics projects
🎨 Aesthetic	Automatic coloring, axes, legends
🚀 Fast	Quick rendering, suitable for real-time applications
📚 Open Source	MIT license, free to use

🚫 Cautions

• Depends on OpenCV: OpenCV must be installed and linked.
• Non-interactive: Cannot zoom or drag charts (suitable for display, not for complex analysis).
• Still under active development: API may change slightly, recommended to follow GitHub.

✅ Conclusion

If you are developing algorithms, robotics, or image processing in C++, and want to quickly visualize data, CVPlot is an excellent choice.

It is like the C++ version of matplotlib.pyplot, simple, direct, and efficient.

📌 One-sentence recommendation:

“With CVPlot, I finally don’t have to draw axes by hand!”