C#轻松实现OpenCV算法零代码开发

OpenCV学堂

发布于 2026-04-02 20:01:07

810

前言介绍

OpenCV实验大师最新版本v1.4.0版本已经支持C#版本工作流引擎集成调用，完成解决实现了OpenCV C#低代码快速开发与应用，成为每个上位机视觉开发人员必须掌握的技能之一。这里给大家详细介绍C#集成与调用OpenCV实验大师工作流引擎SDK的完整步骤与代码演示。

OpenCV安装与环境测试

VS2022 + OpenCVCSharp4.10配置安装与测试

OpenCV官方的开发包至今还不支持C#语言SDK，但是C#有个著名的开源OpenCV Wapper库支持- opencvsharp，github地址如下：

https://github.com/shimat/opencvsharp

用 opencvsharp 就可以跟Python、C++一样轻松在C#语言中使用OpenCV SDK完成各种图像处理与深度学习模型推理部署。

从VS2022中直接安装，截图如下：

代码测试

安装好OpenCVCSharp开发包支持以后，直接创建一个C#语言的控制台空项目，然后创建一个cs文件，把下面的代码直接贴进去

using OpenCvSharp;
class Program 
{
static void Main() 
	{
using var src = new Mat("D:/lena.png", ImreadModes.Grayscale);
using var dst = new Mat();
		Cv2.Canny(src, dst, 50, 200);
using (new Window("src image", src)) 
using (new Window("dst image", dst)) 
		{
			Cv2.WaitKey();
		}
	}
}

OpenCV工作流SDK

下载地址：

https://gitee.com/opencv_ai/opencv_workflow

首先从这里下载OpenCV实验大师工作流SDK开发包，解压缩之后把DLL文件拷贝到对应的C#项目的bin目录里面。

OpenCV工作流引擎SDK演示

首先实现一个多目标多角度多尺度的模板匹配，OpenCV实验大师中导出对应的OpenCV算法工作流文件以后，直接在C#中调用代码如下：

public void run_workflow_ncc_match()
{
    //1.create instance
    IntPtr myInstance = OEMTSWorkflowHelper.workflowClassInit2();
    // 2.load workflow file
    bool succ = OEMTSWorkflowHelper.loadVMConfigFile(myInstance, "D:/match_template_test.vm".ToCharArray(), "ba8e9ea82fb37f1722ad2d9f4962e7e4".ToCharArray());
    // 3.run workflow with input image
    Mat src = Cv2.ImRead("D:/test001.bmp");
    Mat dst = new Mat(src.Size(), src.Type());
    int w = src.Cols;
    int h = src.Rows;
    OEMTSWorkflowHelper.runWorkflow(myInstance, src.CvPtr, dst.CvPtr, w, h);
    int num = OEMTSWorkflowHelper.getNumOfMatchedBoxes(myInstance);
    OEMTSWorkflowHelper.TplBoxInfoWrapper[] tplBoxInfo;
    tplBoxInfo = new OEMTSWorkflowHelper.TplBoxInfoWrapper[num];
    OEMTSWorkflowHelper.getWrapperMatchedBoxInfos(myInstance, tplBoxInfo);
    Console.WriteLine("get num of matched boxes " + num);
    for (int i = 0; i < num; i++)
    {
        Console.WriteLine("box.x: " + tplBoxInfo[i].x);
        Console.WriteLine("box.y: " + tplBoxInfo[i].y);
        Console.WriteLine("box score: " + tplBoxInfo[i].score);
        Console.WriteLine("box angle: " + tplBoxInfo[i].angle);
    }
    // 4.display result image
    Cv2.ImShow("OpenCV学堂-工作流引擎SDK之模板匹配", dst);
    Cv2.ImShow("input", src);
    Cv2.WaitKey(0);
    Cv2.DestroyAllWindows();
    // 5.release resource
    src.Release();
    dst.Release();
    OEMTSWorkflowHelper.deleWorkflowInstance(myInstance);
}

运行结果如下：

再实现一个OpenCV的二值图像分析，OpenCV实验大师中导出对应的OpenCV算法工作流文件以后，直接在C#中调用代码如下：

public void run_ca_workflow()
{
    //1.create instance
    IntPtr myInstance = OEMTSWorkflowHelper.workflowClassInit2();
    // 2.load workflow file
    bool succ = OEMTSWorkflowHelper.loadVMConfigFile(myInstance, "D:/contours_analysis_demo.vm".ToCharArray(), "ba8e9ea82fb37f1722ad2d9f4962e7e4".ToCharArray());
    // 3.run workflow with input image
    Mat src = Cv2.ImRead("D:/2.jpg");
    Mat dst = new Mat(src.Size(), src.Type());
    int w = src.Cols;
    int h = src.Rows;
    OEMTSWorkflowHelper.runWorkflow(myInstance, src.CvPtr, dst.CvPtr, w, h);
    int num = OEMTSWorkflowHelper.getNumOfContours(myInstance);
    Console.WriteLine("get num of contours: " + num);
    // retrieve the contour information
    OEMTSWorkflowHelper.ContourInfoWrapper[] contoursInfo = 
        new OEMTSWorkflowHelper.ContourInfoWrapper[num];
    OEMTSWorkflowHelper.getWrapperContoursInfo(myInstance, contoursInfo);
    for(int i=0; i< contoursInfo.Length; i++)
    {
        Console.WriteLine(i + "th contour info as below");
        Console.WriteLine("roundness: " + contoursInfo[i].roundness);
        Console.WriteLine("contour area: " + contoursInfo[i].area);
    }
    // 4.display result image
    Cv2.ImShow("OpenCV学堂-工作流引擎SDK之轮廓分析", dst);
    Cv2.ImShow("input", src);
    Cv2.WaitKey(0);
    Cv2.DestroyAllWindows();
    // 5.release resource
    src.Release();
    dst.Release();
    OEMTSWorkflowHelper.deleWorkflowInstance(myInstance);
}

运行结果如下：