OpenCV + CPP 系列（卌八）聚类与图像分割（ KMeans）

文章目录

• ​​KMeans- 聚类算法简介​​
• ​​数据聚类代码示例​​
• ​​图像聚类分割​​

KMeans- 聚类算法简介

1. 无监督学习方法
2. 分类问题，输入分类数目，初始化中心位置
3. 硬分类方法，以距离度量
4. 迭代分类为聚类

常见的聚类方法有：划分式聚类方法(​​Partition-based Methods​​)、基于密度的聚类方法(​​Density-based methods​​ DBSCAN)、层次化聚类方法(​​Hierarchical Methods​​) 等。

关于更多理论请查看：​​​​​ 或者查看知乎的大佬：​​https://zhuanlan.zhihu.com/p/104355127​​​0

头文件 ​​machine_learning_all.h​​：

#pragma once
#include <opencv2/opencv.hpp>
#include <iostream>

using namespace cv;
using namespace std;


class Machine_learning{
public:
  void KMeans_data_demo();
};

主函数​​main.cpp​​

#include "machine_learning_all.h"


int main(int argc, char** argv) {
  const char* input_path = "D:\\Desktop\\meinv3.png";
  Mat src = imread(input_path);
  if (src.empty()) {
    cout << "Read image failed!" << endl;
    return -1;
  }

  Machine_learning ml;
  ml.KMeans_data_demo();
  ml.KMeans_image_demo(src);

  imshow("src", src);
  waitKey(0);
  destroyAllWindows();
  return 0;
}

数据聚类代码示例

void Machine_learning::KMeans_data_demo() {
  Mat img = Mat::zeros(450, 450, CV_8UC3);
  RNG rng(123123);

  Scalar colorTab[] = {
    Scalar(0,0,255),
    Scalar(0,255,0),
    Scalar(255,0,0),
    Scalar(0,255,255),
    Scalar(255,0,255),
    Scalar(255,255,0) 
  };

  int numCluster = rng.uniform(3, 5);
  cout << "Number of cluster: " << numCluster << endl;

  int sampleCount = rng.uniform(500, 1000);
  Mat points(sampleCount, 1, CV_32FC2); //生成数据 (height,width)=(sampleCount,1)
  Mat labels;
  Mat centers;
  cout << "points.size():" << points.size() << endl;

  //生成随机数
  for (int k = 0; k < numCluster; k++){
    Point center;
    center.x = rng.uniform(0, img.cols);
    center.y = rng.uniform(0, img.rows);
    Mat PointChunk = points.rowRange(k * sampleCount / numCluster, k == numCluster - 1 ? sampleCount : (k + 1) * sampleCount / numCluster);
    rng.fill(PointChunk, RNG::NORMAL, Scalar(center.x, center.y), Scalar(img.cols * 0.5, img.rows * 0.5));
    randShuffle(points, 1, &rng);
  }

  //聚类
  kmeans(points, numCluster, labels, TermCriteria(TermCriteria::EPS + TermCriteria::COUNT, 10, 0.1), 3, KMEANS_PP_CENTERS, centers);

  //使用不同的颜色区分类别
  img = Scalar::all(0);
  for (int i = 0; i < sampleCount; i++) {
    int index = labels.at<int>(i);
    Point p = points.at<Point2f>(i);
    circle(img, p, 2, colorTab[index], -1, 8, 0);
  }

  //每个聚类的中心来绘图
  for (int i = 0; i < centers.rows; i++) {
    int x = centers.at<float>(i, 0);
    int y = centers.at<float>(i, 1);
    cout << "[x, y] = " << x << ", " << y << endl;
    circle(img, Point(x, y), 40, colorTab[i], 1, LINE_AA);
  }
  imshow("KMeans-demo", img);
}

图像聚类分割

void Machine_learning::KMeans_image_demo(Mat& src) {
  Scalar colorTab[] = {
    Scalar(0,0,255),
    Scalar(0,255,0),
    Scalar(255,0,0),
    Scalar(0,255,255),
    Scalar(255,0,255),
    Scalar(255,255,0)
  };

  int width = src.cols;
  int height = src.rows;
  int dims = src.channels();

  // 初始化定义
  int sampleCount = width * height;
  int clusterCount = 3;
  Mat points(sampleCount, dims, CV_32F, Scalar(10));
  Mat labels;
  Mat centers(clusterCount, 1, points.type());

  // RGB 数据转换到样本数据
  int index = 0;
  for (int row = 0; row < height; row++) {
    Vec3b* bgr_ptr = src.ptr<Vec3b>(row);
    for (int col = 0; col < width; col++) {
      index = row * width + col;
      points.at<float>(index, 0) = static_cast<int>(bgr_ptr[col][0]);
      points.at<float>(index, 1) = static_cast<int>(bgr_ptr[col][1]);
      points.at<float>(index, 2) = static_cast<int>(bgr_ptr[col][2]);
    }
  }

  // 执行K-means聚类
  TermCriteria criteria = TermCriteria(TermCriteria::EPS + TermCriteria::COUNT, 10, 0.1);
  kmeans(points, clusterCount, labels, criteria, 3, KMEANS_PP_CENTERS, centers);

  // 显示图像分割结果
  Mat result = Mat::zeros(src.size(), src.type());
  for (int row = 0; row < height; row++) {
    Vec3b* result_ptr = result.ptr<Vec3b>(row);
    for (int col = 0; col < width; col++) {
      index = row * width + col;
      int label = labels.at<int>(index, 0);
      result_ptr[col][0] = colorTab[label][0];
      result_ptr[col][1] = colorTab[label][1];
      result_ptr[col][2] = colorTab[label][2];
    }
  }
  
  //输出聚类中心
  for (int i = 0; i < centers.rows; i++) {
    int x = centers.at<float>(i, 0);
    int y = centers.at<float>(i, 1);
    cout << "[x, y] = " << x << ", " << y << endl;
    // 聚类中心为：颜色的中心
    //[x, y] = 194, 211
    //[x, y] = 49, 43
    //[x, y] = 146, 149
  }
  imshow("KMeans Image", result);
}