C#结合OpenCVSharp4实现图片相似度识别

using OpenCvSharp;

public Scalar Compare_SSIM(string imgFile1, string imgFile2)
        {
            var image1 = Cv2.ImRead(imgFile1);
            var image2Tmp = Cv2.ImRead(imgFile2);
            // 将两个图片处理成同样大小，否则会有错误： The operation is neither 'array op array' (where arrays have the same size and the same number of channels), nor 'array op scalar', nor 'scalar op array'
            var image2 = new Mat();
            Cv2.Resize(image2Tmp, image2, new OpenCvSharp.Size(image1.Size().Width, image1.Size().Height));
            double C1 = 6.5025, C2 = 58.5225;
            var validImage1 = new Mat();
            var validImage2 = new Mat();
            image1.ConvertTo(validImage1, MatType.CV_32F); //数据类型转换为 float,防止后续计算出现错误
            image2.ConvertTo(validImage2, MatType.CV_32F);
            Mat image1_1 = validImage1.Mul(validImage1); //图像乘积
            Mat image2_2 = validImage2.Mul(validImage2);
            Mat image1_2 = validImage1.Mul(validImage2);
            Mat gausBlur1 = new Mat(), gausBlur2 = new Mat(), gausBlur12 = new Mat();
            Cv2.GaussianBlur(validImage1, gausBlur1, new OpenCvSharp.Size(11, 11), 1.5); //高斯卷积核计算图像均值
            Cv2.GaussianBlur(validImage2, gausBlur2, new OpenCvSharp.Size(11, 11), 1.5);
            Cv2.GaussianBlur(image1_2, gausBlur12, new OpenCvSharp.Size(11, 11), 1.5);
            Mat imageAvgProduct = gausBlur1.Mul(gausBlur2); //均值乘积
            Mat u1Squre = gausBlur1.Mul(gausBlur1); //各自均值的平方
            Mat u2Squre = gausBlur2.Mul(gausBlur2);
            Mat imageConvariance = new Mat(), imageVariance1 = new Mat(), imageVariance2 = new Mat();
            Mat squreAvg1 = new Mat(), squreAvg2 = new Mat();
            Cv2.GaussianBlur(image1_1, squreAvg1, new OpenCvSharp.Size(11, 11), 1.5); //图像平方的均值
            Cv2.GaussianBlur(image2_2, squreAvg2, new OpenCvSharp.Size(11, 11), 1.5);
            imageConvariance = gausBlur12 - gausBlur1.Mul(gausBlur2);// 计算协方差
            imageVariance1 = squreAvg1 - gausBlur1.Mul(gausBlur1); //计算方差
            imageVariance2 = squreAvg2 - gausBlur2.Mul(gausBlur2);
            var member = ((2 * gausBlur1.Mul(gausBlur2) + C1).Mul(2 * imageConvariance + C2));
            var denominator = ((u1Squre + u2Squre + C1).Mul(imageVariance1 + imageVariance2 + C2));
            Mat ssim = new Mat();
            Cv2.Divide(member, denominator, ssim);
            var sclar = Cv2.Mean(ssim);
            return sclar;  // 变化率，即差异
        }