OpenCV学习入门（三）：kmeans原理及代码

<span style="font-size:18px;"><span style="font-size:18px;">// kmeans.cpp : Defines the entry point for the console application.
//

#include "stdafx.h"
#include "opencv2/highgui/highgui.hpp"
#include "opencv2/core/core.hpp"
#include <iostream>

using namespace cv;
using namespace std;

// static void help()
// {
//     cout << "\nThis program demonstrates kmeans clustering.\n"
//             "It generates an image with random points, then assigns a random number of cluster\n"
//             "centers and uses kmeans to move those cluster centers to their representitive location\n"
//             "Call\n"
//             "./kmeans\n" << endl;
// }

int main(int /*argc*/, char** /*argv*/)
{
	const int MAX_CLUSTERS = 5; //最大聚类数目
	Scalar colorTab[] =         //为最大的5种聚类分配5种不同的颜色，用以区分不同类的数据
	{
		Scalar(0, 0, 255),
		Scalar(0, 255, 0),
		Scalar(255, 100, 100),
		Scalar(255, 0, 255),
		Scalar(0, 255, 255)
	};

	Mat img(500, 500, CV_8UC3);  //建立一个500X500，3通道的彩色图
	RNG rng(12345); //随机数产生器
	for (;;)
	{
		int k, clusterCount = rng.uniform(2, MAX_CLUSTERS + 1);		//在[2,6]之间产生均匀随机数作为该次实验聚类个数，保证至少有两个及以上待聚类，才能体现kmeans方法
		int i, sampleCount = rng.uniform(1, 1001);					//在[1,1001]之间均匀随机产生所有样本点
		Mat points(sampleCount, 1, CV_32FC2), labels;				//存放样本点，是sampleCount行2通道的行向量

		clusterCount = MIN(clusterCount, sampleCount);				//保证聚类数小于数据样本数
		Mat centers(clusterCount, 1, points.type());				//用来存储聚类后的中心点

		/* generate random sample from multigaussian distribution */
		for (k = 0; k < clusterCount; k++) 
		{
			Point center;
			center.x = rng.uniform(0, img.cols);					//均匀随机产生初始化聚类中心
			center.y = rng.uniform(0, img.rows);
			Mat pointChunk = points.rowRange(k*sampleCount / clusterCount,	//每个聚类中样本数目都是sampleCount / clusterCount
				k == clusterCount - 1 ? sampleCount :				//K至少能取到0,1，保证有两个及以上待聚类
				(k + 1)*sampleCount / clusterCount);				
			//每个聚类的随机初始化中心作为均值，方差都相等
			rng.fill(pointChunk, CV_RAND_NORMAL, Scalar(center.x, center.y), Scalar(img.cols*0.05, img.rows*0.05));
		}

		randShuffle(points, 1, &rng);								//随机打乱points里面的样本点，注意points和pointChunk是共用数据的。
		//聚类3次，取结果最好的那次，KMEANS_PP_CENTERS表示初始化中心使用Arthur and Vassilvitskii的kmeans++初始化方法。
		kmeans(points, clusterCount, labels,
			TermCriteria(CV_TERMCRIT_EPS + CV_TERMCRIT_ITER, 10, 1.0),
			3, KMEANS_PP_CENTERS, centers);  

		img = Scalar::all(0);
		//遍历所有点，根据不同聚类中心使用不同的颜色绘制出来
		for (i = 0; i < sampleCount; i++)
		{
			int clusterIdx = labels.at<int>(i);
			Point ipt = points.at<Point2f>(i);
			circle(img, ipt, 2, colorTab[clusterIdx], CV_FILLED, CV_AA);
		}

		imshow("clusters", img);

		char key = (char)waitKey();     
		if (key == 27 || key == 'q' || key == 'Q') // 退出条件：按键'ESC','q','Q'
			break;
	}

	return 0;
}

</span></span>