OpenCV5Sharp 1.0.10

OpenCV5Sharp Library

Published by Qourex

Clean, automatic C# wrapper for OpenCV 5.0.0 with complete API coverage and robust unmanaged memory management.

Installation

dotnet add package OpenCV5Sharp

Or via the NuGet Package Manager Console:

Install-Package OpenCV5Sharp

Key Features

• OpenCV 5.0.0 Alignment: Matches the restructured OpenCV 5 API layout (Features module, new Stereo/Calib split, and DNN).
• Auto-Generated Complete API: Over 2,600 methods from core, imgproc, videoio, highgui, features2d, calib3d, objdetect, photo, and dnn modules.
• Thread-Safe Unmanaged Memory: Uses a specialized, atomic Interlocked.Exchange-based DisposableOpenCVObject base class to prevent double-free conditions and memory leaks.
• Exceptions Boundary Protection: Zero raw native crashes. Every single C++ entry point is wrapped in standard try/catch handlers that map exceptions back to managed OpenCVException objects.
• Zero-Configuration Bundling: Prebuilt opencv_world500.dll, opencv_videoio_ffmpeg500_64.dll, and opencv5sharp_native.dll are bundled and copied automatically to the build output.

Supported Platforms

• Operating System: Windows (x64), Linux (x64), macOS (x64, ARM64), Android (ARM64), iOS (ARM64)
• Frameworks: .NET 8.0, .NET 9.0, .NET 10.0

Quick Start Example

using System;
using OpenCV5Sharp;

class Program
{
    static void Main()
    {
        // 1. Read input image (BGR color)
        using (Mat src = Cv2.Imread("input.jpg", (int)ImreadModes.Color))
        using (Mat gray = new Mat())
        using (Mat blurred = new Mat())
        using (Mat edges = new Mat())
        {
            if (src == null || src.Handle == IntPtr.Zero)
            {
                Console.WriteLine("Error: Could not load image.");
                return;
            }

            // 2. Convert to Grayscale
            Cv2.CvtColor(src, gray, (int)ColorConversionCodes.Bgr2gray, 0, AlgorithmHint.Default);

            // 3. Apply a 5x5 Gaussian Blur
            Cv2.GaussianBlur(gray, blurred, new Size(5, 5), 1.5, 1.5, (int)BorderTypes.Default, AlgorithmHint.Default);

            // 4. Perform Canny Edge Detection
            Cv2.Canny(blurred, edges, 50, 150, 3, false);

            // 5. Write the processed image back to disk
            Cv2.Imwrite("output_edges.jpg", edges, IntPtr.Zero);
        } // All unmanaged memory is automatically freed here
    }
}

Type Conversion Guide

Error Handling

All C++ exceptions (including OpenCV's native cv::Exception) are caught on the native side and converted into thread-local error states. When a wrapper call fails:

1. The native function returns a failure status.
2. The managed side detects the failure and fetches the error details.
3. An OpenCVException is thrown containing the native stack trace and error code.

Example:

try
{
    using (Mat src = new Mat()) // Empty image
    using (Mat dst = new Mat())
        {
            Cv2.CvtColor(src, dst, (int)ColorConversionCodes.Bgr2gray, 0, AlgorithmHint.Default);
        }
}
catch (OpenCVException ex)
{
    Console.WriteLine($"OpenCV Error: {ex.Message} (Code: {ex.ErrorCode})");
}

Memory Ownership and Thread Safety

1. Explicit Disposal: Any object inheriting from DisposableOpenCVObject (e.g. Mat, VideoCapture, CLAHE) allocates unmanaged resources. You must dispose them when done, preferably using C# using blocks.
2. Double-Free Prevention: Calling Dispose() multiple times, or having Dispose() race with the Garbage Collector Finalizer, is thread-safe and handled atomically.
3. Thread Safety: P/Invoke calls are thread-safe and re-entrant. However, mutating a single Mat instance from multiple threads simultaneously is not synchronized by the library; you must manage instance-level synchronization in user code.

Support & Contact

• GitHub: github.com/qourex/opencv5sharp
• Website: qourex.com
• Email: info@qourex.com

License

• Library Code: Licensed under the Apache License, Version 2.0.
• FFmpeg Integration: Dynamically links to FFmpeg binaries which are licensed under the GNU LGPL v2.1 or later.
• Trademark Notice: "OpenCV" is a registered trademark of the OpenCV Foundation. This project is a separate community wrapper and is not endorsed by or affiliated with the OpenCV Foundation.

<p align="center">Built with ❤️ by <a href="https://qourex.com">Qourex</a></p>