opencv--FileStorage将矩阵读取写入txt/yml

在使用opencv进行图像处理的过程中，经常会涉及到将文件中的数据读入到cv::Mat中，或者将cv::Mat中的数据写入到txt，yml文件中。

通常是使用FileStorage类来处理，使用流程有3步，如下：

（1）实例化一个FileStorage类对象

写入（FileStorage::WRITE，覆盖写）

追加（FileStorage::APPEND，追加写）

读取（FileStorage::WRITE）

cv::FileStorage fw("calibrateResult.txt", cv::FileStorage::WRITE);

cv::FileStorage fr("calibrateResult.txt", cv::FileStorage::READ);

（2）使用流操作符<<进行文件写入，>>进行文件读取

//文本中的数据名为"cameraMatrix",再将矩阵cameraMatrix写入
fw << "cameraMatrix" << this->cameraMatrix;

（3）使用FileStorage：：release（）函数析构掉类对象，并关闭文件

下面具体看一下代码（2个独立的程序）：

【1】写入参数/矩阵（部分相机标定代码）

这部分把相机标定省略了，因此内外参数没有赋值，你们运行不了，但下一个代码可以。

#include
#include
#include
#include

using namespace std;

int main()
{
	//相机内参矩阵 3x3
	cv::Mat cameraMatrix;
	//畸变参数 1x5
	cv::Mat distCoeffs;
	//保存旋转向量的容器 Mat为3行1列
	vector<cv::Mat> rVecs;
	//保存平移向量的容器 Mat为3行1列
	vector<cv::Mat> tVecs;
	//旋转矩阵
	cv::Mat rMatrix;
	
	//定义字符串
	string rotateVec = "rotVec";
	string rotateMatrix = "rotMat";
	string translate = "traVec";
	//写入文本的话使用WRITE  这里的文本可以是txt yml xml
	cv::FileStorage fw("calibrateResult.txt", cv::FileStorage::WRITE);
	//文本中添加一个"cameraMatrix"的字符串，并将cameraMatrix参数保存到"cameraMatrix"中
	fw << "cameraMatrix" << cameraMatrix;
	fw << "distCoeffs" << distCoeffs;
	
	//我这里有8张图片
	for (int i = 0; i < 8; i++)
	{
		//将旋转向量转换为相对应的旋转矩阵
		cv::Rodrigues(rVecs[i], rMatrix);
		//保存旋转向量
		fw << rotateVec + to_string(i + 1) << rVecs[i];
		//保存旋转矩阵
		fw << rotateMatrix + to_string(i + 1) << rMatrix;
		//保存平移向量
		fw << translate + to_string(i + 1) << tVecs[i];
	}
	//关闭文件
	fw.release();
	
	return 0;
}

【2】读取参数/矩阵（部分相机标定代码）
需要重新建一个工程，并将第1步生成的calibrateResult.txt文件放到与main函数同一目录下，这部分代码你们可以运行

#include
#include
#include
#include

using namespace std;

int main()
{
	//内参
	cv::Mat cameraMatrix;
	//畸变
	cv::Mat distCoeffs;
	//保存旋转矩阵的容器 Mat为3行3列
	vector<cv::Mat> rotMatrix;
	//保存平移向量的容器 Mat为3行1列
	vector<cv::Mat> traMatrix;
	
	//从文本读取参数 这里的文本可以是txt yml xml
	cv::FileStorage fr("calibrateResult.txt", cv::FileStorage::READ);
	if (!fr.isOpened())
		cout << "calibrateResult.txt文件打开失败！" << endl;
	
	//将文本中名字为"cameraMatrix"的数据读给cameraMatrix，读入的时候要加[]
	fr["cameraMatrix"] >> cameraMatrix;
	fr["distCoeffs"] >> distCoeffs;
	
	//定义字符串
	string rName("rotMat");
	string tName("traVec");
	//读入8张图片的旋转矩阵和平移向量
	for (int i = 0; i < 8; i++)
	{
		//定义两个临时变量
		cv::Mat rMat;
		cv::Mat tMat;
		string rotName = rName + to_string(i + 1);
		string traName = tName + to_string(i + 1);
		//将文本中名为"rotMat"的数据读取并赋给rMat
		fr[rotName] >> rMat;
		fr[traName] >> tMat;
		
		//将保存的rMat与tMat通过尾插法存储到名为rotMatrix和traMatrix的容器中
		rotMatrix.push_back(rMat);
		traMatrix.push_back(tMat);
	}
	//关闭文件
	fr.release();
	
	return 0;
}

下面的数据是第1步过程生成的文本文件结果，可以用于第2步的使用。

%YAML:1.0
---
cameraMatrix: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ 5.5656314582396835e+03, 0., 1.2229231977092950e+03, 0.,
       5.5635200465907456e+03, 9.6373914589479955e+02, 0., 0., 1. ]
distCoeffs: !!opencv-matrix
   rows: 1
   cols: 5
   dt: d
   data: [ -1.6449756450459926e-01, 8.6541785695306417e-01,
       3.6332338164954018e-04, -4.7251458379197125e-04,
       -1.1365518697291174e+01 ]
rotVec1: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ 3.7941354063604671e-01, 5.8810276029636809e-02,
       6.8544687767575330e-01 ]
rotMat1: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ 7.7527685213062658e-01, -6.0648151149781304e-01,
       1.7642556153164068e-01, 6.2767053537230333e-01,
       7.0856871266117960e-01, -3.2242840827558122e-01,
       7.0537235385808070e-02, 3.6070840806530519e-01,
       9.3000749608544575e-01 ]
traVec1: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ -2.2242588237238422e+01, -3.0435488494899150e+01,
       2.6680330907120890e+02 ]
rotVec2: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ 1.4961186389998318e-01, 4.7465707169770260e-01,
       2.2423773693015874e+00 ]
rotMat2: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ -6.5693236013450840e-01, -7.0757968802921523e-01,
       2.6032836821326816e-01, 7.5237447046378814e-01,
       -5.9293428104561585e-01, 2.8698709823832291e-01,
       -4.8708627597569887e-02, 3.8439552995501525e-01,
       9.2188260974386727e-01 ]
traVec2: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ 1.4375903305714012e+01, -1.6858126279024283e+01,
       2.6771083226749818e+02 ]
rotVec3: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ 3.6760145441854275e-01, -6.0762638182377604e-01,
       -2.2087152288200689e+00 ]
rotMat3: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ -6.3890998788240005e-01, 6.2727387732740425e-01,
       -4.4533303291666554e-01, -7.6679356773459895e-01,
       -5.6580380618341120e-01, 3.0313969946084229e-01,
       -6.1819510390785093e-02, 5.3515748684944076e-01,
       8.4248727729505446e-01 ]
traVec3: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ 3.6931719283613440e+01, 2.7500637137506775e+01,
       2.6821080189580590e+02 ]
rotVec4: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ 5.4698007983435146e-02, 1.6403368863855264e-01,
       2.1702018511636592e+00 ]
rotMat4: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ -5.6882606064161700e-01, -8.1620386024726777e-01,
       1.0123325171305164e-01, 8.2214726106512137e-01,
       -5.6090516854150951e-01, 9.7258794083310757e-02,
       -2.2600749059678092e-02, 1.3855197732576527e-01,
       9.9009724558805901e-01 ]
traVec4: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ 5.1781917984460499e+01, -9.5653665059897577e+00,
       2.8227136969606414e+02 ]
rotVec5: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ 2.0755534034653123e-01, -3.0226288357576353e-01,
       -2.4420199197937809e+00 ]
rotMat5: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ -7.6986169844529229e-01, 5.9746741208472320e-01,
       -2.2437837855665443e-01, -6.3414379734789494e-01,
       -7.5574810108026069e-01, 1.6342108798674632e-01,
       -7.1934758998148329e-02, 2.6809979337993439e-01,
       9.6070177018547986e-01 ]
traVec5: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ 2.8632431730717109e+01, 3.8421000693274252e+01,
       2.8119574255158761e+02 ]
rotVec6: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ -4.8627107746147763e-01, -2.8177976952282141e-01,
       7.8363942404759201e-01 ]
rotMat6: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ 6.7930675118930872e-01, -6.0434259958218206e-01,
       -4.1630801111542037e-01, 7.3106911467791069e-01,
       6.0667708086959593e-01, 3.1221926447866982e-01,
       6.3877126991454397e-02, -5.1644258333119297e-01,
       8.5393604606523166e-01 ]
traVec6: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ -1.1269422073497639e+01, -3.6890267537566615e+01,
       2.8084310616221831e+02 ]
rotVec7: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ -3.6654883437328739e-01, -1.1909878327701895e-01,
       1.4643363993495367e-01 ]
rotMat7: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ 9.8243724755250850e-01, -1.2079985454024214e-01,
       -1.4221269199515041e-01, 1.6384046806275646e-01,
       9.2319683140578024e-01, 3.4765487413137813e-01,
       8.9293648410332396e-02, -3.6484929166079499e-01,
       9.2677485870527809e-01 ]
traVec7: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ 4.6381644458905322e+00, -1.3983989750510041e+01,
       2.8401280085250306e+02 ]
rotVec8: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ 4.3201085429499669e-01, -1.4205658496125403e-01,
       -2.5761572807677602e-01 ]
rotMat8: !!opencv-matrix
   rows: 3
   cols: 3
   dt: d
   data: [ 9.5770321158010518e-01, 2.1605290307018885e-01,
       -1.9006762376611064e-01, -2.7603846540120142e-01,
       8.7635404039009557e-01, -3.9472314539548403e-01,
       8.1285448563085047e-02, 4.3049359921710595e-01,
       8.9892599077176527e-01 ]
traVec8: !!opencv-matrix
   rows: 3
   cols: 1
   dt: d
   data: [ -2.2614457056945270e+01, 4.0665059749313768e+00,
       2.6709282364848053e+02 ]