本文通过摄像头参数（fx,fy,cx,cy,k1,k2,p1,p2,p3（标定得到））去矫正摄像头拍出来的图像畸变

详细代码在底部

首先

这里我们先介绍两个函数：他们都可以用来矫正畸变，但是一个是输入是Mat型，一个输入是IplImage* 型

本文用Mat格式读取图像，故采用第一个函数。

undistort()        //用于Mat图像 cvUndistort()     //用于 IplImage* 图像

其次

标定一下摄像头，得到参数矩阵。（我用matlab标定工具箱搞出来的~）

Matlab2019 标定工具箱结果

再次

这里通过paramtersInit()将参数转换为系数阵

然后传入到undistort里面去，这就结束了~

效果

畸变矫正前/后

代码

```

#include <iostream>

#include<opencv2/core/core.hpp>

#include <opencv2/highgui/highgui.hpp>

#include <opencv2/imgproc/imgproc.hpp>

using namespace cv;

using namespace std;

Mat matrix;

Mat coeff;

Mat new_matrix;

double fx = 682.421509, fy = 682.421509, cx = 633.947449, cy = 404.559906;

double k1k2Distortion[2] = {-0.15656,-0.00404};

double p1p2p3Distortion[3] = {-0.00078,-0.00048,0.00000};

using namespace cv;

using namespace std;

void paramtersInit();//将参数转换为系数阵

int main()

{

cv::Mat srcMat, dstMat;

paramtersInit();

srcMat = imread("img.jpg");

undistort(srcMat, dstMat, matrix, coeff, new_matrix);

imwrite("undistortImg.png", dstMat);

imshow("front", srcMat);

imshow("after",dstMat);

waitKey(0);

system("pause");

return 0;

}

void paramtersInit()

{

matrix = Mat(3, 3, CV_64F);

matrix.at<double>(0, 0) = fx;

matrix.at<double>(0, 1) = 0;

matrix.at<double>(0, 2) = cx;

matrix.at<double>(1, 0) = 0;

matrix.at<double>(1, 1) = fy;

matrix.at<double>(1, 2) = cy;

matrix.at<double>(2, 0) = 0;

matrix.at<double>(2, 1) = 0;

matrix.at<double>(2, 2) = 1;

coeff = Mat(1, 4, CV_64F);

coeff.at<double>(0, 0) = k1k2Distortion[0];

coeff.at<double>(0, 1) = k1k2Distortion[1];

coeff.at<double>(0, 2) = p1p2p3Distortion[0];

coeff.at<double>(0, 3) = p1p2p3Distortion[1];

}

```