Use of OpenCV 11 kmeans clustering

Program running logic

Before clustering , use first

Blur(src, src, new Size(15, 15));

The mean filter is applied to the graph

For clustering , The following steps are carried out

1、var columnVector = src.Reshape(cn: 3, rows: src.Rows * src.Cols);

Turn the graph into 3 Single column matrix of channels

2、columnVector.ConvertTo(samples, MatType.CV_32FC3);

Convert a matrix to a floating point matrix

3、 call kmean function , And look at the different parameters k：2,4,5,8 The effect of

Cv2.Kmeans(
                    data: samples,
                    k: clustersCount,
                    bestLabels: bestLabels,
                    criteria:
                        new TermCriteria(type: CriteriaType.Eps | CriteriaType.MaxIter, maxCount: 10, epsilon: 1.0),
                    attempts: 3,
                    flags: KMeansFlags.PpCenters,
                    centers: centers);

4、 Read generation matrix bestLabel, And rewrite the matrix of the same type as the original drawing

var clusteredImage = new Mat(src.Rows, src.Cols, src.Type());
                for (var size = 0; size < src.Cols * src.Rows; size++)
                {
                    var clusterIndex = bestLabels.At<int>(0, size);
                    var newPixel = new Vec3b
                    {
                        Item0 = (byte)(centers.At<float>(clusterIndex, 0)), // B
                        Item1 = (byte)(centers.At<float>(clusterIndex, 1)), // G
                        Item2 = (byte)(centers.At<float>(clusterIndex, 2)) // R
                    };
                    clusteredImage.Set(size / src.Cols, size % src.Cols, newPixel);
                }

5、 graphic display .