数字图像处理（二值化，高通，低通滤波，中值滤波，线性灰度拉伸，缩放，平移旋转）OpenCV

川长思鸟来 2022-11-13 05:26 219阅读 0赞

1.二值化

#include "stdafx.h"
    #include <opencv2/core.hpp>
    #include <opencv2/imgcodecs.hpp>
    #include <opencv2/highgui.hpp>
    #include <iostream>
    #include <math.h>
    using namespace cv;
    
    Mat Judge(Mat input, int h, int w, int rmin, int max);
    Mat State(Mat src);
    //获取图像最大灰度级和最小灰度级
    
    int main()
    {
        
    	std::string image_path = "bj.bmp";
    	//读入原灰度图像
    	Mat img_source = imread(image_path, IMREAD_GRAYSCALE);
    	//如果为空的话提示打不开
    	if (img_source.empty())
    	{
        
    		std::cout << "Could not read the image: " << image_path << std::endl;
    		return 1;
    	}
    	//定义判断法和状态法的图像
    	Mat Anal;
    	Mat Sta;
    	int rmin = 0;
    	int rmax = 255;
    	int h{
         img_source.rows };				// 图片高
    	int w{
         img_source.cols };               // 图片宽度
    	namedWindow("【原图】");
    	imshow("【原图】", img_source);
    	Anal = Judge(img_source, h, w, rmin, rmax);
    	namedWindow("【判断分析图】");
    	imshow("【判断分析图】", img_source);
    	img_source = imread(image_path, IMREAD_GRAYSCALE);
    	Sta = State(img_source);
    	namedWindow("【状态图】");
    	imshow("【状态图】", img_source);
    	int k = waitKey(1055500); // Wait for a keystroke in the window
    
    	if (k == 'J')
    	{
        
    		//保存图片
    		imwrite("Judge_0.bmp", Anal);
    
    	}
    	else
    		imwrite("State_0.bmp", Sta);
    	return 0;
    }
    
    Mat Judge(Mat input, int h, int w, int rmin, int rmax)
    {
        //参考实习教材
    	int T = 0;
    	int hist[256];
    	unsigned char* ptr = input.data;
    	memset(hist, 0, sizeof(hist));
    	int S, S1, S2;
    	double P, P1;
    	double mean1, mean2;
    	double D1, D2, classinDis, classoutDis;
    	double Tmax;
    	int graymin = 255, graymax = 0;
    	int i, j;
    	int k = 0;
    	//统计灰度直方图，各个灰度级个数
    	for (i = 0; i<h; i++)
    		for (j = 0; j < w; j++)
    		{
        
    			int gray = ptr[i*w + j];
    			hist[gray]++;
    			if (gray > graymax) graymax = gray;
    			if (gray < graymin) graymin = gray;
    		}
    	P = 0.0;
    	S = 0;
    	if (graymin == 0)
    		graymin++;
    	//计算总的质量矩P和像素总数S
    	for (k = graymin; k <= graymax; k++)
    	{
        
    		P += (double)k*(double)hist[k];
    		S += hist[k];
    	}
    	Tmax = 0.0;
    	S1 = 0, S2 = 0;
    	P1 = 0.0;
    	D1 = 0.0;
    	D2 = 0.0;
    	classinDis = 0.0;
    	classoutDis = 0.0;
    	double ratio = 0.0;
    	//求阈值
    	for (k = graymin; k <= graymax; k++)
    	{
        //计算1类像素的个数
    		S1 += hist[k];
    		if (!S1)
    		{
        
    			continue;
    		}
    		//计算2类像素的个数
    		S2 = S - S1;
    		if (S2 == 0)
    		{
        
    			break;
    		}
    		//计算1类质量矩
    		P1 += (double)k*hist[k];
    		//计算1类均值
    		mean1 = P1 / S1;
    		//计算1类间方差
    		for (int n = graymin; n <= k; n++)
    		{
        
    			D1 += ((n - mean1)*(n - mean1)*hist[n]);
    		}
    		
    		//同上
    		mean2 = (P - P1) / S2;
    		for (int m = k + 1; m <= graymax; m++)
    		{
        
    			D2 += ((m - mean2)*(m - mean2)*hist[m]);
    		}
    		classinDis = D1 + D2;
    		classoutDis = (double)S1*(double)S2*(mean1 - mean2)*(mean1 - mean2);
    		if (classinDis != 0)
    			ratio = classoutDis / classinDis;
    		if (ratio >Tmax)
    		{
        
    			Tmax = ratio;
    			T = k;
    		}
    	}
    	//二值化
    	for (i = 0; i < h; i++)
    		for (j = 0; j < w; j++) {
        
    			if (ptr[i*w + j] < T)
    				ptr[i*w + j] = 0;
    			else  ptr[i*w + j] = 255;
    		}
    	return input;
    }
    
    
    Mat State(Mat src)
    {
        //参考教程
    	int iThrehold;//阀值
    	int height = src.rows;
    	int width = src.cols;
    	unsigned char *ptr = src.data;
    
    	int iDiffRec = 0;
    	int hist[256] = {
         0 }; //直方图数组  
    	int iTotalGray = 0;//灰度值和  
    	int iTotalPixel = 0;//像素数和  
    	int gray;//某点的像素值  
    
    	int iNewThrehold;//新阀值
    	int iMaxGrayValue = 0, iMinGrayValue = 255;//原图像中的最大灰度值和最小灰度值  
    	int iMeanGrayValue1, iMeanGrayValue2;
    
    	//获取(i,j)的值，存于直方图数组F  
    	for (int i = 0; i < width; i++)
    	{
        
    		for (int j = 0; j < height; j++)
    		{
        
    			gray = ptr[j*width + i];
    			if (gray < iMinGrayValue)
    				iMinGrayValue = gray;
    			if (gray> iMaxGrayValue)
    				iMaxGrayValue = gray;
    			hist[gray]++;
    		}
    	}
    
    	iThrehold = 0;
    	iNewThrehold = (iMinGrayValue + iMaxGrayValue) / 2;//初始阀值  
    	iDiffRec = iMaxGrayValue - iMinGrayValue;
    
    	for (int a = 0; (abs(iThrehold - iNewThrehold) > 0.5); a++)//迭代中止条件  
    	{
        
    		iThrehold = iNewThrehold;
    		//小于当前阀值部分的平均灰度值  
    		for (int i = iMinGrayValue; i < iThrehold; i++)
    		{
        
    			iTotalGray += hist[i] * i;//hist[]存储图像信息  
    			iTotalPixel += hist[i];
    		}
    		iMeanGrayValue1 = (double)(iTotalGray / iTotalPixel);
    		//大于当前阀值部分的平均灰度值  
    		iTotalPixel = 0;
    		iTotalGray = 0;
    		for (int j = iThrehold + 1; j < iMaxGrayValue; j++)
    		{
        
    			iTotalGray += hist[j] * j;//hist[]存储图像信息  
    			iTotalPixel += hist[j];
    		}
    		iMeanGrayValue2 = (double)(iTotalGray / iTotalPixel);
    
    		iNewThrehold = (iMeanGrayValue2 + iMeanGrayValue1) / 2; //新阀值  
    		iDiffRec = abs(iMeanGrayValue2 - iMeanGrayValue1);
    	}
    	//二值化
    	for (int i = 0; i < height; i++)
    		for (int j = 0; j < width; j++) {
        
    			if (ptr[i*width + j] < iThrehold)
    				ptr[i*width + j] = 0;
    			else  ptr[i*width + j] = 255;
    		}
    	return src;
    }

结果图：  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70]  
2.线性拉伸

// Linear _stretch.cpp : 定义控制台应用程序的入口点。
    //
    
    #include "stdafx.h"
    #include <opencv2/core.hpp>
    #include <opencv2/imgcodecs.hpp>
    #include <opencv2/highgui.hpp>
    #include <iostream>
    #include <math.h>
    using namespace cv;
    //获取图像最大灰度级和最小灰度级
    int RminRmax(Mat input, int &rmin, int& rmax);
    /*
    //图像对比度拉伸。
    //（r1,s1)灰度平面上的第一个点
    //（r2,s2)灰度平面上的第二个点
    */
    Mat Stretch(Mat input, int r1, int s1, int r2, int s2, int rmin, int rmax);
    int InsertSort(int arr[], int size);
    
    int main()
    {
        
    	std::string image_path = "bj.bmp";
    	//读入原灰度图像
    	Mat img_source = imread(image_path, IMREAD_GRAYSCALE);
    	//如果为空的话提示打不开
    	if (img_source.empty())
    	{
        
    		std::cout << "Could not read the image: " << image_path << std::endl;
    		return 1;
    	}
    	int h{
         img_source.rows };				// 图片高
    	int w{
         img_source.cols };               // 图片宽度
    	namedWindow("【原图】");
    	imshow("【原图】", img_source);
    	Mat img_pro;//拉伸处理后的图像
    	int rmin = 0;
    	int rmax = 0;
    	RminRmax(img_source, rmin, rmax);
    	//h灰度拉伸
    	img_pro = Stretch(img_source, 105, 40, 230, 220,rmin,rmax);
    	namedWindow("【线性变换图】");
    	imshow("【线性变换图】", img_pro);	
    	int k = waitKey(1055500); // Wait for a keystroke in the window
    
    	if (k == 'w')
    	{
        
    		//保存图片
    		 imwrite("1_0.bmp", img_pro);
    
    	}
        return 0;
    }
    int RminRmax(Mat input, int &rmin, int &rmax)
    {
        
    	int h{
         input.rows };				// 图片高
    	int w{
         input.cols };				// 图片宽
    	uchar* data_in{
         input.data };		// 原图图片像素数据
    
    	for (int y = 0; y < h; y++)
    		for (int x = 0; x < w; x++)
    		{
        
    			if (rmin>data_in[y*w + x])
    				rmin = data_in[y*w + x];
    
    			if (rmax < data_in[y*w + x])
    				rmax = data_in[y*w + x];
    
    		}
    
    	return 0;
    
    }
    //线性变化
    Mat Stretch(Mat input, int r1, int s1, int r2, int s2, int rmin, int rmax)
    {
        //(r1s1) (r2,s2)
    	int h{
         input.rows };				// 图片高
    	int w{
         input.cols };               // 图片宽度
    	unsigned char* ptr = input.data;
    
    	double data;            //灰度
    	double k, k1, k2, k3, b, b1, b2, b3;		// 直线斜率与截距
    
    	k1 = (double)(s1 - rmin) / (double)(r1 - rmin);
    	b1 = 0;
    	k2 = (double)(s2 - s1) / (double)(r2 - r1);
    	b2 = s1 - k2*r1;
    	k3 = (rmax - s2) / (rmax - r2);
    	b3 = s2 - k3*r2;
    
    
    	for (int y = 0; y < h; y++)
    		for (int x = 0; x < w; x++)
    		{
        
    			data = ptr[y*w + x];
    			if (data <= r1)
    			{
        
    				k = k1;
    				b = b1;
    			}
    			else if (data >= r2)
    			{
        
    				k = k3;
    				b = b3;
    			}
    			else
    			{
        
    				k = k2;
    				b = b2;
    			}
    			data = k*data + b;
    
    			ptr[y*w + x] = (data);
    		}
    
    	return input;
    }
    
    //获取图片灰度值最大值与最小值
    int RmimRmax(Mat input, int &rmin, int& rmax)
    {
        
    	int h{
         input.rows };				// 图片高
    	int w{
         input.cols };				// 图片宽
    	uchar* data_in{
         input.data };		// 原图图片像素数据
    
    	for (int y = 0; y < h; y++)
    		for (int x = 0; x < w; x++)
    		{
        
    			if (rmin>data_in[y*w + x])
    				rmin = data_in[y*w + x];
    
    			if (rmax < data_in[y*w + x])
    				rmax = data_in[y*w + x];
    
    		}
    	return 0;
    
    }

![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 1]  
3.高低通，中值滤波

#include "stdafx.h"
    #include <opencv2/core.hpp>
    #include <opencv2/imgcodecs.hpp>
    #include <opencv2/highgui.hpp>
    #include <iostream>
    #include <math.h>
    //像素点的位置数组
    int dx[] = {
         -1,0,1,-1,0,1,-1,0,1 };
    int dy[] = {
         -1,-1,-1,0,0,0,1,1,1 };
    using namespace cv;
    //中值滤波
    Mat image_filter_median(Mat img_source, Mat dst);
    int InsertSort(int arr[], int size);
    //高低通
    Mat HighPass(Mat src, int h, int w);
    Mat LowPass(Mat src, int h, int w);
    int main()
    {
        
    	std::string image_path = "bj.bmp";
    	//读入原灰度图像
    	Mat img_source = imread(image_path, IMREAD_GRAYSCALE);
    	//如果为空的话提示打不开
    	if (img_source.empty())
    	{
        
    		std::cout << "Could not read the image: " << image_path << std::endl;
    		return 1;
    	}
    	int h{
         img_source.rows };				// 图片高
    	int w{
         img_source.cols };               // 图片宽度
    	namedWindow("【原图】");
    	imshow("【原图】", img_source);
    	Mat dst;//中值滤波
        dst.create(img_source.size(), img_source.type());
    	dst=image_filter_median(img_source, dst); 
    	namedWindow("【中值滤波图】");
    	imshow("【中值滤波图】", dst);
    	Mat Low;//低通滤波图像
    	Mat High;
    	
    	High=HighPass(img_source, h, w);
    	namedWindow("【高通滤波图】");
    	imshow("【高通滤波图】", High);
    	Low=LowPass(img_source, h, w);
    	namedWindow("【低通滤波图】");
    	imshow("【低通滤波图】", Low);
    
    	int k = waitKey(1055500); // Wait for a keystroke in the window
    	if (k == 'm')
    	{
        
    		//保存图片
    		 imwrite("2_0.bmp", dst);
    
    	}
    	else if (k == 'H')
    	{
        
    		//保存图片
    		imwrite("3_0.bmp",High );
    	}
    	else  imwrite("4_0.bmp", Low);
    	return 0;
    }
    //插入排序，返回中间值
    int InsertSort(int arr[], int size)
    {
        
    	for (int i = 1; i < size; i++)
    	{
        
    		int temp = arr[i];
    		int j = i - 1;
    		while (j >= 0)
    		{
        
    			if (arr[j] > temp)arr[j + 1] = arr[j];
    			else break;
    			j--;
    		}
    		arr[j + 1] = temp;
    	}
    	return arr[4];
    }
    //中值滤波
    Mat image_filter_median(Mat img_source, Mat dst)
    {
        
    	for (int i = 0; i <img_source.rows; i++) {
        
    		for (int j = 0; j < img_source.cols; j++) {
        
    			if ((i - 1) >= 0 && (i + 1) <img_source.rows && (j - 1) >= 0 && (j + 1) < img_source.cols)
    			{
        
    				int d[9];
    				for (int k = 0; k < 9; ++k) {
        
    					int gray = img_source.at<uchar>(i + dx[k], j + dy[k]);
    					d[k] = gray;
    				}
    				dst.at<uchar>(i, j) = InsertSort(d, 9);
    			}
    			else {
        
    				int gray = img_source.at<uchar>(i, j);
    				dst.at<uchar>(i, j) = gray;
    			}
    		}
    	}
    	return dst;
    	
    }
    
    //高通滤波
    Mat HighPass(Mat src, int h,int w)
    {
        
       int i, j;
       unsigned char *ptr = src.data;
       Mat ResultImage(h, w, src.type());
       unsigned char *ptr2 = ResultImage.data;
       unsigned char *p1, *p2,  *p3, *p4, *p5, *p6, *p7, *p8, *p9 ;
       for(int i=1;i<h-1;i++)
    	   for (int j = 1; j < w - 1; j++)
    	   {
        
    		   p1 = &ptr[ w*(i - 1) + (j - 1) ];
    		   p2 = &ptr[w*(i ) + (j - 1)];
    		   p3= &ptr[w*(i + 1) + (j - 1)];
    		   p4 = &ptr[w*(i - 1) + (j)];
    		   p5 =&ptr2[w*(i ) + (j ) ];
    		   p6 = &ptr[w*(i + 1) + (j )];
    		   p7 = &ptr[w*(i - 1) + (j + 1)];
    		   p8 = &ptr[w*(i) + (j + 1)];
    		   p9 = &ptr[w*(i+1) + (j+ 1)];
    		   *p5=ptr[ w*(i)+(j)];
    		   int t = (int)(*p5 * 9 - *p1 - *p2 - *p3 - *p4 - *p5 - *p6 - *p7 - *p8 - *p9);
    		   if (t > 255)
    			   *p5 = 255;
    		   else if (t < 0)
    			   *p5 = 0;
    		   else *p5 = t;
    		   ptr2[w*(i-1) + j-1] = *p5;
    	   }
       return ResultImage;
    
    }
    Mat LowPass(Mat src, int h, int w)
    {
        
    	int i, j;
    	unsigned char *ptr = src.data;
    	Mat ResultImage(h, w, src.type());
    	unsigned char *ptr2 = ResultImage.data;
    	unsigned char *p1, *p2, *p3, *p4, *p5, *p6, *p7, *p8, *p9;
    	for (int i = 1; i<h - 1; i++)
    		for (int j = 1; j < w - 1; j++)
    		{
        
    			p1 = &ptr[w*(i - 1) + (j - 1)];
    			p2 = &ptr[w*(i)+(j - 1)];
    			p3 = &ptr[w*(i + 1) + (j - 1)];
    			p4 = &ptr[w*(i - 1) + (j)];
    			p5 = &ptr2[w*(i)+(j)];
    			p6 = &ptr[w*(i + 1) + (j)];
    			p7 = &ptr[w*(i - 1) + (j + 1)];
    			p8 = &ptr[w*(i)+(j + 1)];
    			p9 = &ptr[w*(i + 1) + (j + 1)];
    			*p5 = ptr[w*(i)+(j)];
    			int t = (int)( (*p1 + *p2 + *p3 + *p4 + *p5 +*p6+ *p7 + *p8 + *p9)/9);
    			if (t > 255)
    				*p5 = 255;
    			else if (t < 0)
    				*p5 = 0;
    			else *p5 = t;
    			ptr2[w*(i - 1) + j - 1] = *p5;
    		}
    	return ResultImage;
    	
    }

![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 2]  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 3]  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 4]  
4.缩放，平移，旋转

// Translation _rotation.cpp : 定义控制台应用程序的入口点。
    //
    #include "stdafx.h"
    #include <opencv2/core.hpp>
    #include <opencv2/imgcodecs.hpp>
    #include <opencv2/highgui.hpp>
    #include <iostream>
    #include <math.h>
    using namespace cv;
    Mat Translation(Mat src, int h, int w, int nX, int nY);
    Mat Shrink(Mat src, float scale, int h, int w);
    Mat Rotate(Mat src, float angle, int h, int w);
    
    int main()
    {
        
    	std::string image_path = "bj.bmp";
    	//读入原灰度图像
    	Mat img_source = imread(image_path, IMREAD_GRAYSCALE);
    	Mat Trans;
    	Mat Shrink1;
    	Mat Rotate1;
    	//如果为空的话提示打不开
    	if (img_source.empty())
    	{
        
    		std::cout << "Could not read the image: " << image_path << std::endl;
    		return 1;
    	}
    	int h{
         img_source.rows };				// 图片高
    	int w{
         img_source.cols };               // 图片宽度
    	namedWindow("【原图】");
    	imshow("【原图】", img_source);
    	Trans=Translation(img_source, h, w, 35, 50);
    	namedWindow("【平移】");
    	imshow("【平移】", Trans);
    	Shrink1=Shrink(img_source, 0.5, h, w);
    	namedWindow("【缩放】");
    	imshow("【缩放】", Shrink1);
    
    	Rotate1=Rotate(img_source, 30, h, w);
    	namedWindow("【旋转】");
    	imshow("【旋转】",Rotate1);
    	int k = waitKey(1055500); // Wait for a keystroke in the window
    
    	if (k == 'T')
    	{
        
    		//保存图片
    		 imwrite("T_0.bmp", Trans);
    
    	}
    	else if(k=='S')
    		imwrite("S_0.bmp", Shrink1);
    	else  imwrite("R_0.bmp", Rotate1);
        return 0;
    }
    //平移
    Mat Translation(Mat src, int h, int w, int nX, int nY)
    {
        
    	Mat ResultImage(h, w, src.type());
    	unsigned char *ptr = src.data;
    	unsigned char *ptr2 = ResultImage.data;
    
    	for (int i = 0; i < h; i++)
    	{
        
    		//y是变换后的坐标=当前行号+变换Y
    		int y = (i + nY);
    		for (int j = 0; j < w; j++)
    		{
        
    			//x是变换后的坐标=当前列号+变换X
    			int x = (j + nX);
    			if (y< h && x<w)
    			{
        //行列坐标要位于（w，h）内
    			 //交换位置
    				ptr2[x + y*w] = ptr[j + i*w];
    			}
    
    		}
    	}
    	return ResultImage;
    
    }
    
    Mat Shrink(Mat src, float scale, int h, int w)
    {
        //scale是缩小的比例
     //高度和宽度放缩后的大小为m1,m2
    	int m1 = floor(h*scale);
    	int n1 = floor(w*scale);
    	Mat ResultImage = Mat::zeros(m1, n1, src.type());
    	unsigned char *ptr = src.data;
    	unsigned char *ptr2 = ResultImage.data;
    	//分别得到x,y的变化
    	float ScaleY = float(m1) / float(h);
    	float ScaleX = float(n1) / float(w);
    	//分配内存
    	for (int i = 0; i < m1; i++)
    	{
        
    		//变化后的坐标
    		float InvY = float(i) / ScaleY;
    		float Y = floor(InvY);
    		//变换了多少
    		float Dy = InvY - Y;
    		for (int j = 0; j < n1; j++)
    		{
        
    			float InvX = float(j) / ScaleX;
    			float X = floor(InvX);
    			float Dx = InvX - X;
    			//取整
    			int x = int(X);
    			int y = int(Y);
    			ptr2[j + i*n1] = (1 - Dy)*(1 - Dx)*ptr[x + y*w] + Dy*(1 - Dx)*ptr[x + (y + 1)*w]
    				+ (1 - Dy)*Dx*ptr[x + 1 + y *w]
    				+ Dy*Dx*ptr[x + 1 + (y + 1) *w];
    		}
    	}
    	Mat ResultImage1;
    	ResultImage.convertTo(ResultImage1, CV_8UC1);
    	return ResultImage;
    
    }
    //旋转
    float *Sin_Cos(float x,float y, float sin, float cos)
    {
        
    	float n[2] = {
        0};
    	n[0] = cos*x + sin*y;
    	n[1] = -sin*x + cos*y;
    	return n;
    }
    
    Mat Rotate(Mat src, float angle, int h, int w)
    {
        
    	Mat ResultImage = Mat::zeros(h, w, src.type());
    	unsigned char *ptr = src.data;
    	
    	float SinAngle, CosAngle;
    	//旋转后的图像宽高
    	int newH, newW;
    	int i0, j0;
    	//图像四个角的坐标/变换后的坐标
    	int SrcX[4] = {
         0 };
    	int SrcY[4] = {
         0 };
    	int DstX[4] = {
         0 };
    	int DstY[4] = {
         0 };
    	//角度转为弧度
    	float RotateAngle = angle*3.1415926 / 180;
    	SinAngle = (float)(sin((double)RotateAngle));
    	CosAngle = (float)(cos((double)RotateAngle));
    	SrcX[0] = (float)(-(w - 1) / 2);SrcY[0] = (float)((h - 1) / 2);
    	SrcX[1] = (float)((w - 1) / 2);SrcY[1] = (float)((h - 1) / 2);
    	SrcX[2] = (float)(-(w - 1) / 2);SrcY[2] = (float)(-(h - 1) / 2);
    	SrcX[3] = (float)((w - 1) / 2);SrcY[3] = (float)(-(h - 1) / 2);
    	
    	for (int i = 0; i < 4; i++)
    	{
        
    		float *n = Sin_Cos(SrcX[i], SrcY[i], SinAngle, CosAngle);
    		DstX[i] = n[0];
    		DstY[i] = n[1];
    	}
    	float f1, f2;
    	newW = (long)(max(fabs(DstX[3] - DstX[0]), fabs(DstX[2] - DstX[1] + 0.5)));
    	newH=(long)(max(fabs(DstY[3] - DstY[0]), fabs(DstY[2] - DstY[1] + 0.5)));
    	f1 = (float)(-0.5*(newW - 1)*CosAngle - 0.5*(newH - 1)*SinAngle + 0.5*(newW - 1));
    	f2 = (float)(0.5*(newW - 1)*SinAngle - 0.5*(newH - 1)*CosAngle + 0.5*(newH - 1));
    	int i, j;
    	ResultImage = Mat::zeros(newH, newW, src.type());
    	unsigned char *ptr2 = ResultImage.data;
    	for (i = 0; i <newH; i++)
    	{
        
    		for (j = 0; j <newW; j++)
    		{
        
    			i0 = (-((float)j))*SinAngle + ((float)i)*CosAngle + f2 + 0.5;
    			j0 = (((float)j))*CosAngle + ((float)i)*SinAngle + f1 + 0.5;
    			if ((j0 >= 0) && (j0 < w) && (i0 >= 0) && (i0 < h))
    				ptr2[i*newW + j]=ptr[i0*w + j0];
    			else ptr2[i*newW + j] = 0;
    		}
    	}
    
    	return ResultImage;
    
    }

![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 5]  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 6]

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70]: /images/20221022/aa09aed8c62c43baaa7c5d4e5f0260f5.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 1]: /images/20221022/4f87d67696d943469c80cc72901ace59.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 2]: /images/20221022/af2ea13bd6714689af2f8071c7960c69.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 3]: /images/20221022/d4956c91c70f477e99413b744ab4e077.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 4]: /images/20221022/071ebf42559e4d8b80355a4d8902500e.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 5]: /images/20221022/a2fe5584b855451e994e81f15fed8fd8.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ2ODc3Njk3_size_16_color_FFFFFF_t_70 6]: /images/20221022/4cfd0a0a97124685bd673504c8c310db.png