msr图像增强opencv c++ 实现
时间: 2023-10-25 13:10:17 浏览: 41
MSR图像增强是一种基于多尺度Retinex理论的图像增强方法,可以有效地提高图像的对比度和亮度,并且不会造成颜色失真。下面是使用OpenCV C++实现MSR图像增强的步骤:
1.加载图像并转换为灰度图像:
```c++
Mat src = imread("image.jpg");
cvtColor(src, src, CV_BGR2GRAY);
```
2.将图像的像素值转换为浮点类型,并将像素值缩放到0到1之间:
```c++
src.convertTo(src, CV_32F, 1.0 / 255);
```
3.计算图像的对数值:
```c++
Mat log_src;
log(src + 0.01, log_src);
```
4.分解图像为高频和低频成分:
```c++
Mat gauss;
GaussianBlur(log_src, gauss, Size(0, 0), 10);
Mat low_freq = log_src - gauss;
```
5.将低频成分进行缩放:
```c++
normalize(low_freq, low_freq, 0, 1, NORM_MINMAX);
```
6.计算低频成分的对数值:
```c++
Mat log_low_freq;
log(low_freq + 0.01, log_low_freq);
```
7.计算高频成分:
```c++
Mat high_freq = log_src - log_low_freq;
```
8.将高频成分进行缩放:
```c++
normalize(high_freq, high_freq, 0, 1, NORM_MINMAX);
```
9.将高频成分和低频成分加权融合:
```c++
Mat dst = high_freq * 0.5 + low_freq * 0.5;
```
10.将图像像素值重新缩放到0到255之间:
```c++
dst.convertTo(dst, CV_8U, 255);
```
完整代码如下:
```c++
#include <opencv2/opencv.hpp>
using namespace cv;
using namespace std;
int main()
{
Mat src = imread("image.jpg");
cvtColor(src, src, CV_BGR2GRAY);
src.convertTo(src, CV_32F, 1.0 / 255);
Mat log_src;
log(src + 0.01, log_src);
Mat gauss;
GaussianBlur(log_src, gauss, Size(0, 0), 10);
Mat low_freq = log_src - gauss;
normalize(low_freq, low_freq, 0, 1, NORM_MINMAX);
Mat log_low_freq;
log(low_freq + 0.01, log_low_freq);
Mat high_freq = log_src - log_low_freq;
normalize(high_freq, high_freq, 0, 1, NORM_MINMAX);
Mat dst = high_freq * 0.5 + low_freq * 0.5;
dst.convertTo(dst, CV_8U, 255);
imshow("src", src);
imshow("dst", dst);
waitKey(0);
return 0;
}
```