数字图像处理基础实验（二）：直方图均衡化
一、实验内容及原理
1、计算灰度图像的归一化直方图
具体内容：利用 OpenCV 对图像像素进行操作，计算归一化直方图.并在 窗口中以图形的方式显示出来
步骤：
（1）初始化一个256大小的数组，存放图像中每级灰度值的像素点的个数
（2）遍历整张图像
（3）归一化数组
2、灰度图像直方图均衡处理
具体内容：通过计算归一化直方图，设计算法实现直方图均衡化处理。
步骤：
（1）读取该像素点的灰度值
（2）计算其在归一化直方图中的累加概率
（3）累加概率乘以255，并四舍五入
3、彩色图像直方图均衡处理
具体内容： 在灰度图像直方图均衡处理的基础上实现彩色直方图均衡处理
步骤：
（1）读取该像素点的三通道灰度值
（2）计算其在归一化直方图中的累加概率
（3）累加概率乘以255，并四舍五入
二、实验代码
实验环境：
（1）OpenCV3.4.3
（2）Ubuntu16.04
（3）VS Code
（4）C++
#include
#include
#include
#include
#include
class Experiment2 {
public:
Experiment2(std::vector path){
for(int i = 0; i < path.size(); i++){
original_color_image.push_back(cv::imread(path[i]));
original_gray_image.push_back(color2Gray(original_color_image[i]));
}
}
// 0.1、彩色图像转灰度图像
cv::Mat color2Gray(cv::Mat src_image){
//创建与原图同类型和同大小的矩阵
cv::Mat gray_image(src_image.rows, src_image.cols, CV_8UC1);
if(src_image.channels()!=1){
for(int i = 0; i < src_image.rows; i++)
for(int j = 0; j < src_image.cols; j++)
gray_image.at(i, j) = (src_image.at(i, j)[0] + src_image.at(i, j)[1] + src_image.at(i, j)[2]) / 3;
}
else
gray_image = src_image.clone();
return gray_image;
}
// 2.1、计算灰度图像的归一化直方图
void computerGrayHistogram(int pic_id, int opt=1){
std::vector hist(256,0);
cv::Mat image = (opt > 0) ? original_gray_image[pic_id] : gray_image_equalization[pic_id];
float nums = image.rows * image.cols;
float max = 0;
for(int i = 0; i < image.rows; i++)
for(int j = 0; j < image.cols; j++)
hist[image.at(i, j)]++;
for(int i=0; i < 256; i++){
hist[i] = hist[i] / nums;
if(hist[i] > max)
max = hist[i];
}
if(opt)
gray_hist.push_back(hist);
else
gray_hist_equalization.push_back(hist);
drawGrayHist(max, pic_id, opt);
}
// 2.2、绘制归一化直方图
void drawGrayHist(float maxValue, int pic_id, int opt=1){
int scale = 20000;
int hist_h = scale * maxValue, hist_w = 257;
std::vector hist = (opt > 0) ? gray_hist[pic_id] : gray_hist_equalization[pic_id];
// 画图底色
cv::Mat image_hist = cv::Mat::zeros(hist_h, hist_w, CV_8U);
// 画出直方图
for (int i = 0; i < 256; i++){
float binValue = hist[i];
int realValue = cv::saturate_cast(binValue * scale);
cv::rectangle(image_hist,
cv::Point(i, hist_h - realValue),
cv::Point(i + 1, hist_h),
cv::Scalar(255));
}
if(opt)
gray_image_hist.push_back(image_hist);
else
gray_image_hist_equalization.push_back(image_hist);
}
// 3、灰度图像直方图均衡处理
void grayEqualization(int pic_id){
cv::Mat image_equalization = cv::Mat::zeros(original_gray_image[pic_id].size(), original_gray_image[pic_id].type());
for(int i = 0; i < original_gray_image[pic_id].rows; i++)
for(int j = 0; j < original_gray_image[pic_id].cols; j++){
int k = original_gray_image[pic_id].at(i, j);
float sum = 0;
for(int m = 0; m <= k; m++)
sum += gray_hist[pic_id][m];
image_equalization.at(i, j) = sum * 255 + 0.5;
}
gray_image_equalization.push_back(image_equalization);
}
// 4.1、彩色图像直方图均衡处理
void colorEqualization(int pic_id){
cv::Mat image_equalization = cv::Mat::zeros(original_color_image[pic_id].size(), original_color_image[pic_id].type());
for(int i = 0; i < original_color_image[pic_id].rows; i++)
for(int j = 0; j < original_color_image[pic_id].cols; j++)
for(int n = 0; n < 3; n++){
int k = original_color_image[pic_id].at(i, j)[n];
float sum = 0;
for(int m = 0; m <= k; m++)
sum += color_hist[pic_id][n][m];
image_equalization.at(i, j)[n] = sum * 255 + 0.5;
}
color_image_equalization.push_back(image_equalization);
}
// 4.2、计算彩色图像的归一化直方图
void computerColorHistogram(int pic_id){
std::vector > hist(3, std::vector(256));
int nums = original_color_image[pic_id].rows * original_color_image[pic_id].cols;
for(int i = 0; i < original_color_image[pic_id].rows; i++)
for(int j = 0; j < original_color_image[pic_id].cols; j++){
hist[0][original_color_image[pic_id].at(i, j)[0]]++;
hist[1][original_color_image[pic_id].at(i, j)[1]]++;
hist[2][original_color_image[pic_id].at(i, j)[2]]++;
}
for(int i=0; i < 256; i++){
hist[0][i] = hist[0][i] / nums;
hist[1][i] = hist[1][i] / nums;
hist[2][i] = hist[2][i] / nums;
}
color_hist.push_back(hist);
}
// 5 显示 灰度均衡化 与 灰度直方图
void test_GrayEqualizationAndHistogram(){
for(int i = 0; i < original_gray_image.size(); i++){
computerGrayHistogram(i);
grayEqualization(i);
std::cout<< "灰度图像: 当前已经均衡化第" << i + 1 << "张图像\n";
}
// 显示结果
cv::imshow(" 原始灰度图像", original_gray_image[0]);
cv::waitKey(0);
for(int i = 0; i < original_gray_image.size(); i++){
cv::imshow("原始图像的灰度直方图:", gray_image_hist[i]);
cv::waitKey(0);
cv::imshow("均衡化后的灰度图像:", gray_image_equalization[i]);
cv::waitKey(0);
}
std::cout<< "\n";
}
// 6 显示 彩色均衡化
void test_ColorEqualizationAndHistogram(){
for(int i = 0; i < original_color_image.size(); i++){
computerColorHistogram(i);
colorEqualization(i);
std::cout<< "彩色图像: 当前已经均衡化第" << i + 1 << "张图像\n";
}
// 显示结果
cv::imshow("原始图像", original_color_image[0]);
cv::waitKey(0);
for(int i = 0; i < original_color_image.size(); i++){
cv::imshow("均衡化后---彩色图像:", color_image_equalization[i]);
cv::waitKey(0);
}
}
private:
std::vector color_info;
std::vector gray_info;
// 原始图像
std::vector original_color_image;
std::vector original_gray_image;
// 均衡化后的图像
std::vector color_image_equalization;
std::vector gray_image_equalization;
// 画出来的灰度直方图
std::vector gray_image_hist;
std::vector > color_image_hist;
std::vector gray_image_hist_equalization;
std::vector > color_image_hist_equalization;
// 灰度图像的灰度分布 (图像名,每级灰度的个数)
std::vector > gray_hist;
std::vector > gray_hist_equalization;
// 彩色图像的灰度分布(图像名,通道,每级灰度的个数)
std::vector > > color_hist;
std::vector > > color_hist_equalization;
};
int main(){
std::vector path;
path.push_back("/home/lyd/image_process/pic/lena.jpg");
Experiment2 a(path);
a.test_GrayEqualizationAndHistogram();
a.test_ColorEqualizationAndHistogram();
return 1;
}
三、实验结果
