【图像处理 直方图 OpenCV实现】

直方图与直方图匹配

最后一个代码比较实用，对于小角度的模板匹配，还是可以用此函数的。

---

前言

直方图统计的原理、直方图代表的图像特征、直方图的应用。

直方图作为图像处理的基础，是分析图像的基础工具。不可不了解

直方图是对像素数据统计的一种方法。它可以图示数据分布，可以描绘图像中每个亮度值的像素数。

直方图的作用：
1：对像素数据进行统计。这将可以了解整个图像的像素分布。进而分析图像的亮暗对比。
2：借助直方图来实现图像的二值化。

直方图的操作
1：归一化 normalize
直方图统计完成以后，我们需要对其进行进一步分析。倘若一个图片像素在某个区间内很均匀，但是在另一个区间内分布值很高或者很低。那么整体的直方图可能就不能显示出重要特征的灰度分布。为了可以灵活展现灰度分布，查看像素的占用比例，我们引用归一化的方式进行统计完后的处理。
归一化有很多种方式。normalize();中第五个参数给大家提供了5种归一化的方式。1：：无穷范数
2：：L1范数，绝对值之和 4：：L2范数及模长归一化 5：：L2范数平方 32：：偏移归一化

2：直方图比较
3：直方图均衡
4：直方图匹配
5：直方图反向投影
6：图像的模板匹配

在Opencv中，提供了图像直方图的统计函数calcHist()

---

一、直方图

二、使用步骤

1.引入库

绘制直方图（示例）：

#include
#include

using namespace std;
using namespace cv;

2.绘制直方图

绘制直方图（示例）：

int main1() 
{
	Mat img;
	img = imread("apple.jpg");

	if (img.empty())
	{
		cout << "图像数据有问题" << endl;
		return -1;
	}

	Mat gray;
	cvtColor(img,gray,COLOR_BGR2GRAY);

//以下为提取直方图的必要变量
	Mat hist;
	const int channels[1] = { 0 };//通道索引
	float inRanges[2] = { 0,255 };//每个通道灰度值的取值范围
	const float*ranges[1] = { inRanges };//像素的灰度值范围
	const int bins[1] = { 256 };//每个直方图数组的尺寸
	
	calcHist(&gray,1,channels,Mat(),hist,1,bins,ranges);  //统计函数
	//准备绘制直方图
	int hist_w = 512;//对应直方图的列，因绘制的宽度是2，即需要256*2
	int hist_h = 400;//对应行，若不进行归一化，则需要大于255
	int width = 2;//绘制的间隙

	Mat histImage = Mat::zeros(hist_h,hist_w,CV_8UC3);
	for (int i = 1; i <= hist.rows; i++)
	{
	//简介：从图像的左下角开始绘制，即第一个点的坐标是0，hist_h-1。后面每个增width。  取整然后除以15是为了缩小范围，类似归一。
		rectangle(histImage,Point(width*(i-1),hist_h-1),Point(width*i-1,hist_h-cvRound(hist.at<float>(i-1)/15)),Scalar(255,255,255),-1);
		cout << hist.at<float>(i - 1) << endl;
	}
	namedWindow("直方图",WINDOW_NORMAL);
	imshow("直方图",histImage);

	waitKey(0);
	return 0;
}

//归一化
int main2() 
{
	system("color F0");

	vector<double> positiveData = { 2.0,8.0,10.0 };//为了只管观察归一化，定义了一个数组
	vector<double> normalized_l1, normalized_L2, normalized_Inf, normalized_L2SQR;

	

	//测试不同的归一方法
	normalize(positiveData,normalized_l1,1.0,0.0,NORM_L1);
	cout << "normalized_L1" << normalized_l1[0] << ":" << normalized_l1[1] << ":" << normalized_l1[2] << ";;;;"<<endl;
	normalize(positiveData, normalized_L2, 1.0, 0.0, NORM_L2);
	cout << "normalized_L2" << normalized_L2[0] << "+" << normalized_L2[1] << "+" << normalized_L2[2] << endl;
	normalize(positiveData,normalized_Inf,1.0,0.0,NORM_INF);
	cout << "normalized_inf" << normalized_Inf[0] << "+" << normalized_Inf[1] << "+"  << normalized_Inf[2] << endl;
	normalize(positiveData, normalized_L2SQR, 1.0, 0.0, NORM_MINMAX);
	cout << "MINMAX" << normalized_L2SQR[0] << normalized_L2SQR[1] << normalized_L2SQR[2] << endl;

	//将图像直方图归一化
	//Mat img = imread("apple.jpg");

	Mat img(512, 512, CV_32FC3, Scalar(100, 100, 100));
	if (img.empty())
	{
		cout << "图像数据有问题" << endl;
		return -1;
	}
	imshow("apple",img);

	Mat gray, hist;
	cvtColor(img, gray, COLOR_BGR2GRAY);
	const int channels[1] = { 0 };
	float inRanges[2] = {0,255};
	const float*ranges[1] = { inRanges };
	const int bins[1] = { 256 };

	calcHist(&gray,1,channels,Mat(),hist,1,bins,ranges);

	int hist_w = 512;
	int hist_h = 400;
	int width = 2;


	Mat hisImage_L1 = Mat::zeros(hist_h,hist_w,CV_8UC3);
	Mat hisImage_Inf = Mat::zeros(hist_h, hist_w, CV_8UC3);

	Mat hist_L1, hist_Inf;

	normalize(hist,hist_L1,1,0,NORM_L1,-1,Mat());
	//hist_L1是一个行为256，列为1的矩阵。所以遍历了rows。每个矩阵的没一个点，代表了这个像素归一化的值。
	for (int i = 1; i < hist_L1.rows; i++)
	{
		rectangle(hisImage_L1,Point(width*(i-1),hist_h-1),Point(width*i-1,hist_h-cvRound(30*hist_h*hist_L1.at<float>(i-1))-1),Scalar(255,255,255),-1);

	}

	normalize(hist,hist_Inf,1,0,NORM_INF,-1,Mat());

	for (int i = 1; i < hist_Inf.rows; i++)
	{
		rectangle(hisImage_Inf,Point(width*(i-1),hist_h-1),Point(width*i-1,hist_h-cvRound(hist_h*hist_Inf.at<float>(i-1))-1),Scalar(255,255,255),-1);
		cout << hist_Inf.at<float>(i - 1) << endl;
	}
	
	namedWindow("L1归一化",WINDOW_NORMAL);
	imshow("L1归一化",hisImage_L1);

	namedWindow("INF归一化",WINDOW_NORMAL);
	imshow("INF归一化",hisImage_Inf);

	waitKey(0);
	return 0;


}

//直方图的反向投影
void drawHist3(Mat &hist,int type,string name) 
{
	int hist_w = 512;
	int hist_h = 400;
	int width = 2;
	Mat histImage = Mat::zeros(hist_h,hist_w,CV_8UC3);
	normalize(hist,hist,255,0,type,-1,Mat());
	namedWindow(name,WINDOW_NORMAL);
	imshow(name,hist);
}

int main66() 
{
	Mat img0 = imread("apple.jpg");
	Mat img1 = imread("sub.jpg");

	Mat img0_HSV, img1_HSV, hist0, hist1;
	if (img0.empty()||img1.empty())
	{
		cout << "图像数据有问题" << endl;
		waitKey(3000);
		return - 1;
	}

	cvtColor(img1,img1_HSV,COLOR_BGR2HSV);
	cvtColor(img0, img0_HSV, COLOR_BGR2HSV);
	int h_bins = 32; int s_bins = 32;
	int histSize[] = { h_bins,s_bins };

	//H通道为0-179
	float h_ranges[] = { 0,180 };
	//S通道
	float s_ranges[] = { 0,256 };
	//每个通道的范围
	const float*ranges[] = { h_ranges,s_ranges };
	//统计通道的索引
	int channels[] = { 0,1 };
	//绘制H-S 二维直方图
	//calcHist(&img0_HSV,1,channels,Mat(),hist0,2,histSize,ranges,true,false);
	calcHist(&img1_HSV, 1, channels, Mat(), hist0, 2, histSize, ranges, true, false);

	drawHist3(img0_HSV,NORM_INF,"hist00");
	//drawHist3(img1_HSV,NORM_INF,"hist1");

	double dur;
	clock_t start, end;
	start = clock();


	imwrite("111.png",img0);
	end = clock();
	dur = (double)(end - start);

	printf("Use Time:%f\n", (dur / CLOCKS_PER_SEC));


	Mat backproj;



	calcBackProject(&img0_HSV,1,channels,hist0,backproj,ranges,1.0);

	imshow("反向投影的结果",backproj);
	waitKey(0);
	return 0;



}

//比较直方图
void drawingHist(Mat &hist,int type ,string name) 
{
	int width = 2;
	int hist_h = 400;
	int hist_w = 512;
	Mat histImage = Mat::zeros(hist_h,hist_w,CV_8UC3);
	normalize(hist,hist,1,0,type,-1,Mat());

	for (int i = 1; i < hist.rows; i++)
	{
		rectangle(histImage,Point(width*(i-1),hist_h-1),Point(width*(i-1),hist_h-cvRound(hist_h*hist.at<float>(i-1))-1),Scalar(255,255,255),-1);
	}
	imshow(name, histImage);

}


int main55() 
{

	Mat img = imread("apple.jpg");
	if (img.empty())
	{
		cout << "图像数据有问题" << endl;
		return -1;
	}
	Mat gray, gray2, gray3, hist, hist2, hist3;
	cvtColor(img, gray,COLOR_BGR2GRAY);
	resize(gray,gray2,Size(),0.5,0.5);
	gray3 = imread("lena.png",IMREAD_GRAYSCALE);

	const int channels[1] = { 0 };
	float inRanges[2] = { 0,255 };
	const float*ranges[1] = { inRanges };
	const int bins[1] = { 256 };
	calcHist(&gray, 1, channels, Mat(), hist, 1, bins, ranges);
	calcHist(&gray2,1, channels, Mat(), hist2,1,bins,ranges);
	calcHist(&gray3,1, channels, Mat(), hist3,1,bins,ranges);

	drawingHist(hist,NORM_INF,"hist");
	drawingHist(hist2,NORM_INF,"hist2");
	drawingHist(hist3,NORM_INF,"hist3");

	//将原来的直方图和现在的直方图进行比较
	double histNumer = compareHist(hist,hist2,HISTCMP_CORREL);
	cout << histNumer << endl;

	double histN2 = compareHist(hist,hist3,HISTCMP_CORREL);
	cout << histN2 << endl;

	waitKey(0);
	return 0;
}

//直方图均衡
void  drawHist(Mat &hist,int type,string name)  //归一化并绘制直方图函数 
{
	int hist_w = 600;
	int hist_h = 400;
	int width = 2;
	Mat histImage = Mat::zeros(hist_h,hist_w,CV_8UC3);
	normalize(hist,hist,1,0,type,-1,Mat());

	for (int i = 1; i < hist.rows; i++)
	{
		rectangle(histImage,Point(width*(i-1),hist_h-1),Point(width*i-1,hist_h-cvRound(hist_h*hist.at<float>(i-1))-1),Scalar(255,255,255),-1);
	}
	imshow(name,histImage);
}

int main4() 
{
	Mat img = imread("gearwHeel.jpg");
	if (img.empty())
	{
		cout << "数据有问题" << endl;
		return -1;
	}
	Mat gray, hist, hist2;
	cvtColor(img,gray,COLOR_BGR2GRAY);
	Mat equalImg;
	equalizeHist(gray,equalImg);//将直方图均值化

	const int channels[1] = { 0 };
	float inRanges[2] = { 0,255 };
	const float*ranges[1] = {inRanges};
	const int bins[1] = { 256 };


	calcHist(&gray,1,channels,Mat(),hist,1,bins,ranges);
	calcHist(&equalImg,1,channels,Mat(),hist2,1,bins,ranges);

	drawHist(hist,NORM_INF,"hist");
	drawHist(hist2,NORM_INF,"hist2");

	namedWindow("原图",WINDOW_NORMAL);
	imshow("原图",gray);
	imshow("均值化图",equalImg);

	waitKey(0);
	return 0;

}

//图像的模板匹配
int main() 
{
	Mat img = imread("4.Bmp");
	Mat temp = imread("6.Bmp");//待匹配的图，可以是从原图截取出来的一张图

	if (img.empty() || temp.empty())
	{
		cout << "图像数据有问题" << endl;

		waitKey(3000);
		return -1;

	}

	Mat result;
	matchTemplate(img,temp,result,TM_CCOEFF_NORMED);

	double maxVal, minVal;
	Point minLoc, maxLoc;
	//寻找匹配结果最大的值
	minMaxLoc(result,&minVal,&maxVal,&minLoc,&maxLoc);
	//绘制最佳的矩形
	rectangle(img,cv::Rect(maxLoc.x,maxLoc.y,temp.cols,temp.rows),Scalar(0,0,0),8);
	namedWindow("原图",WINDOW_NORMAL);
	imshow("原图",img);
	namedWindow("匹配图",WINDOW_NORMAL);
	imshow("匹配图",temp);
	namedWindow("result",WINDOW_AUTOSIZE);
	imshow("result",result);

	waitKey(0);
	return 0;
}

该处使用的url网络请求的数据。

---

总结

代码参考 《OpenCV 4快速入门》 冯振 郭延宁 吕跃勇

本门意在总结在本书学到的知识、复习。

亲测对于简单的，角度差异小的模板匹配，灰度匹配的效果还是可以的。

后期会抽时间进一步讲解此文的代码。以上代码欢迎交流