图像处理与计算机视觉-图像处理简单操作及解析

#include<iostream> #include<opencv2\opencv.hpp> #include"generateGs.h"; #include"hist.h"; using namespace std; using namespace cv; //extern void generateGaussianTemplate(double window[3][3], int ksize, double sigma); const int model_size = 1;//模板大小 const int model_element_number = (model_size * 2 + 1)*(model_size * 2 + 1);//模板元素的个数 const int model_centre_loction = (model_element_number + 1) / 2;//模板中心元素的位置 double model_element[model_element_number] = { 0.1111,0.1111,0.1111,0.111,0.1111,0.111,0.111,0.111,0.1111 };//模板的元素，数组有多大就写多少元素。 double model_gs[][3] = { {0.1,0.3,0.1},{ 0.05,0.2,0.05 },{0.05,0.1,0.05}};//手工设计的高斯模板 //double model_lpls[][3] = { { 0,-1,0 },{ -1,4,-1 },{ 0,-1,0 } };//常用的拉普拉斯卷积核 //double model_lpls[][3] = { { 1,1,1 },{ 1,-8,1 },{ 1,1,1 } };//常用的拉普拉斯卷积核 double model_lpls[][3] = { { -1,-1,-1 },{ -1,8,-1 },{ -1,-1,-1 } };//常用的拉普拉斯卷积核 double sigma = 0.6;//高斯模板生成时方差的赋值 int main(int argc, char **argv) { Mat srcImage = imread("C://Users//Administrator//Desktop//2.bmp");//读入图像 int cols = srcImage.cols, rows = srcImage.rows;//获取图像大小 Mat grayImage = Mat::zeros(rows, cols, CV_8UC1); Mat dstImage = Mat::zeros(rows, cols, CV_8UC1);//设置相同大小的灰度图 Mat gsImage = Mat::zeros(rows, cols, CV_8UC1);//设置相同大小的灰度图 Mat histImage = Mat::zeros(rows, cols, CV_8UC1);//设置相同大小的灰度图 Mat lplsImage = Mat::zeros(rows, cols, CV_8UC1);//设置相同大小的灰度图 cvtColor(srcImage, grayImage, CV_BGR2GRAY);//转化为灰度图 imshow("灰度图像", grayImage); //下面是遍历图像元素 double sum = 0;//均值模板与对应像素的卷积和 double sumGs = 0;//高斯模板与对应像素的卷积和 int arry_i = 0;//均值滤波数组下标 int arry_gi = 0;//高斯滤波数组下标 int arry_gj = 0;//高斯滤波数组下标 int arry_li = 0;//拉普拉斯滤波数组下标 int arry_lj = 0;//拉普拉斯滤波数组下标 for (int i = 0 + model_size; i < rows - model_size; i++) { for (int j = 0 + model_size; j < cols - model_size; j++) { //以下是基于循环的均值滤波 sum = 0; arry_i = 0;//数组下标清0 //遍历模板元素并与对应像素卷积 for (int k = i - model_size; k <= i + model_size; k++) { for (int m = j - model_size; m <= j + model_size; m++) { sum += (double)grayImage.at<uchar>(k, m)*model_element[arry_i];//模板元素与对应像素相乘，然后求和； arry_i += 1; } } dstImage.at<uchar>(i, j) = (uchar)sum; } } //生成高斯模板,这个模板生成的不太好，自己设计的更理想，就用上边自己定义的观察 //generateGs(model_gs,model_element_number,sigma); //高斯滤波 for (int i = 0 + model_size; i < rows - model_size; i++) { for (int j = 0 + model_size; j < cols - model_size; j++) { //以下是基于循环的高斯滤波 sum = 0; arry_gi = 0;//数组下标清0 arry_gj = 0;//数组下标清0 //遍历模板元素并与对应像素卷积 for (int k = i - model_size; k < i + model_size; k++) { for (int m = j - model_size; m < j + model_size; m++) { sum += (double)grayImage.at<uchar>(k, m)*model_gs[arry_gi][arry_gj];//模板元素与对应像素相乘，然后求和； arry_gj += 1; } arry_gj = 0; arry_gi += 1; } gsImage.at<uchar>(i, j) = (uchar)sum; } } //直方图均衡化 histImage = equalize_hist(grayImage); //下边注释用于检验错误，要确保模板中数相加=1 /*double a = 0; for (int ii = 0; ii < 3; ii++) { for (int jj = 0; jj < 3; jj++) { a += model_gs[ii][jj]; } } cout << a << endl*/; //拉普拉斯滤波器进行图像锐化 for (int i = 0 + model_size; i < rows - model_size; i++) { for (int j = 0 + model_size; j < cols - model_size; j++) { //以下是基于循环的高斯滤波 sum = 0; arry_li = 0;//数组下标清0 arry_lj = 0;//数组下标清0 //遍历模板元素并与对应像素卷积 for (int k = i - model_size; k < i + model_size; k++) { for (int m = j - model_size; m < j + model_size; m++) { sum += (double)grayImage.at<uchar>(k, m)*model_lpls[arry_li][arry_lj];//模板元素与对应像素相乘，然后求和； arry_lj += 1; } arry_lj = 0; arry_li += 1; } lplsImage.at<uchar>(i, j) = (uchar)sum; } } namedWindow("均值滤波后图像"); imshow("均值滤波后图像", dstImage); imshow("高斯滤波后图像", gsImage); imshow("直方图均衡化后图像", histImage); imshow("拉普拉斯滤波后图像", lplsImage); waitKey(0); return 0; }