c++ opencv数字图像处理：低通滤波之统计排序滤波器（中值滤波器）

---

前言

数字图像处理c++ opencv（VS2019 opencv4.53）持续更新

---

一、中值滤波器

中值滤波器即取领域像素值的中值作为像素结果

1.代码

代码如下（示例）：

#include<iostream>
#include<opencv2/opencv.hpp>
#include"Salt.h"

using namespace cv;
using namespace std;

void myfilter(int filter_size, Mat& image_input, Mat& image_output);
int mysort( Mat& sortarray, int filter_size);

int main()
{ 
	Mat image, image_gray, image_output, image_output2;   //定义输入图像，灰度图像，输出图像
	image = imread("lena.png");  //读取图像；
	if (image.empty())
	{
		cout << "读取错误" << endl;
		return -1;
	}
	imshow("image", image);

	cvtColor(image, image_gray, COLOR_BGR2GRAY);
	Salt(image_gray, 1000);   //添加噪声
	imshow("image_gray", image_gray);

	//opencv自带滤波函数
	int a = 7;
	medianBlur(image_gray, image_output,a);
	imshow("image_output", image_output);

	//自编函数
    myfilter(a, image_gray, image_output2);
    imshow("image_output2", image_output2);


	waitKey(0);  //暂停，保持图像显示，等待按键结束
	return 0;
}


void myfilter(int filter_size, Mat& image_input, Mat& image_output) //图像边缘未处理
{
	image_output = image_input.clone();
	Mat sortarray(1, filter_size* filter_size, CV_8U);
	int k = (filter_size - 1) / 2;

	for (int i = k; i < (image_input.rows-k); i++)
	{
		for (int j = k; j < (image_input.cols - k); j++)
		{
			int h = 0;
			for (int m = -k; m < k + 1; m++)
			{
				for (int n = -k; n < k + 1; n++)
				{
					sortarray.at<uchar>(h) = image_input.at<uchar>(i + m, j + n);
					h++;
				}
			}

			image_output.at<uchar>(i, j) = mysort(sortarray, filter_size* filter_size);

		}
	}
}

int mysort(Mat& sortarray, int filter_size)  //找中值
{
	int exchange;
	int m;
	for (int i = 1; i <= filter_size - 1; i++)
	{
		for (int j = 1; j <= filter_size - i; j++)
		{
			if (sortarray.at<uchar>(j - 1) > sortarray.at<uchar>(j))
			{
				exchange = sortarray.at<uchar>(j - 1);
				sortarray.at<uchar>(j - 1) = sortarray.at<uchar>(j);
				sortarray.at<uchar>(j) = exchange;
			}
		}

	}
	return m = sortarray.at<uchar>((filter_size + 1) / 2 );
}

结果：