To say with AI What does it matter , It is estimated that only the library itself , however , This big map search small map function is still good , You can try something you like . My main scene is , A specific interface is a big picture , An icon in the interface is a small figure , perhaps , The whole desktop is a big picture , A small icon is a small diagram to determine the location of a plug-in or function .

I also wrote before based on C# Code mapping , But I don't feel efficient , This time, we will make a decision based on OpenCvSharp Map finding function of , If you like, you can try .

Create a Winfrom project

Because the project example is simple , however , You also need to get the desktop size , We have to create such a project .

You want to add Nuget package

Install-Package OpenCvSharp4.Windows -Version 4.6.0.20220608

Find the main business code

public static Rectangle Find(Image sourceImage, Image matchImage, double threshold = 0.8)
{
    var refMat = Mat.FromImageData(ImageHelper.ImageToBytes(sourceImage), ImreadModes.AnyColor);// Big picture 
    var tplMat = Mat.FromImageData(ImageHelper.ImageToBytes(matchImage), ImreadModes.AnyColor);// Little picture 
    using (Mat res = new Mat(refMat.Rows - tplMat.Rows + 1, refMat.Cols - tplMat.Cols + 1, MatType.CV_32FC1))
    {
        Mat gref = refMat.CvtColor(ColorConversionCodes.BGR2GRAY);
        Mat gtpl = tplMat.CvtColor(ColorConversionCodes.BGR2GRAY);

        Cv2.MatchTemplate(gref, gtpl, res, TemplateMatchModes.CCoeffNormed);
        Cv2.Threshold(res, res, 0.8, 1.0, ThresholdTypes.Tozero);

        double minval, maxval;
        Point minloc, maxloc;

        Cv2.MinMaxLoc(res, out minval, out maxval, out minloc, out maxloc);

        if (maxval >= threshold)
        {
            return new Rectangle(maxloc.X, maxloc.Y, tplMat.Width, tplMat.Height);
        }
        return Rectangle.Empty;
    }
}

Search can be realized according to similarity , In fact, if you take a screenshot directly , Similarity is 0.99, It's pretty good .

If the multi graph result returns , The boss who wants to study , You can study .

effect

You can see that the search speed is still very fast , Of course , The first time may be a little slower , near 200 The ms

summary

This case is solved , It means going WEB Page slide map unlock verification Automation is another step forward , Think about how to do this case later .

Code address

https://github.com/kesshei/SearchImageDemo.git

https://gitee.com/kesshei/SearchImageDemo.git

reading

One button three times ！, Thank you for your support , Your support is my motivation !