ORB-SLAM3的源码学习： Settings.cc：Settings::readCamera1/readCamera2 从配置文件中加载相机参数

PaLu-LI2025-02-18 0:35

前言

需要从配置文件yaml文件中读取相机参数才能用于后续计算。

1.函数声明

读取相机1的参数：

cpp 复制代码 
void Settings::readCamera1(cv::FileStorage &fSettings)

如果是双目相机则还要读取相机2的参数：

cpp 复制代码 
void Settings::readCamera2(cv::FileStorage &fSettings)

2.函数定义

相机1

1.读取相机模型

3的模型加入了针孔相机模型以及广角相机模型，可以利用模板函数读取参数，来确定相机的类型。

cpp 复制代码 
 bool found;

    // Read camera model
    string cameraModel = readParameter<string>(fSettings, "Camera.type", found);
2.如果是针孔相机模型

就进行如下操作：

读取相机参数构造相机模型，针对不同畸变参数的个数采用不同的方式进行读取，如果是单目相机和RGBD相机的话就需要进行去畸变操作。

cpp 复制代码 
vector<float> vCalibration;
    if (cameraModel == "PinHole")
    {
        cameraType_ = PinHole;//设置相机类型

        // Read intrinsic parameters读取内参
        float fx = readParameter<float>(fSettings, "Camera1.fx", found);
        float fy = readParameter<float>(fSettings, "Camera1.fy", found);
        float cx = readParameter<float>(fSettings, "Camera1.cx", found);
        float cy = readParameter<float>(fSettings, "Camera1.cy", found);

        vCalibration = {fx, fy, cx, cy};//储存在校准的变量中

        calibration1_ = new Pinhole(vCalibration);
        originalCalib1_ = new Pinhole(vCalibration);

        // Check if it is a distorted PinHole
        // 检查这个针孔相机的配置文件是否有畸变参数
        readParameter<float>(fSettings, "Camera1.k1", found, false);
        if (found)
        {
            readParameter<float>(fSettings, "Camera1.k3", found, false);
            if (found)
            {
                vPinHoleDistorsion1_.resize(5);//储存相机畸变参数的向量扩容。
                vPinHoleDistorsion1_[4] = readParameter<float>(fSettings, "Camera1.k3", found);
            }
            else
            {
                vPinHoleDistorsion1_.resize(4);
            }
            vPinHoleDistorsion1_[0] = readParameter<float>(fSettings, "Camera1.k1", found);
            vPinHoleDistorsion1_[1] = readParameter<float>(fSettings, "Camera1.k2", found);
            vPinHoleDistorsion1_[2] = readParameter<float>(fSettings, "Camera1.p1", found);
            vPinHoleDistorsion1_[3] = readParameter<float>(fSettings, "Camera1.p2", found);
        }

        // Check if we need to correct distortion from the images
        // 如果是单目相机或者是RGBD相机且有畸变参数则需要去畸变操作。
        if (
            (sensor_ == System::MONOCULAR ||
                sensor_ == System::IMU_MONOCULAR ||
                sensor_ == System::RGBD ||
                sensor_ == System::IMU_RGBD) &&
            vPinHoleDistorsion1_.size() != 0)
        {
            bNeedToUndistort_ = true;
        }
    }
3.如果是矫正后的图像

此时认为是没有畸变的，就正常读取参数构造相机模型即可。

cpp 复制代码 
 //如果是校正后的图像则认为没畸变。
    else if (cameraModel == "Rectified")
    {
        cameraType_ = Rectified;

        // Read intrinsic parameters
        float fx = readParameter<float>(fSettings, "Camera1.fx", found);
        float fy = readParameter<float>(fSettings, "Camera1.fy", found);
        float cx = readParameter<float>(fSettings, "Camera1.cx", found);
        float cy = readParameter<float>(fSettings, "Camera1.cy", found);

        vCalibration = {fx, fy, cx, cy};

        calibration1_ = new Pinhole(vCalibration);
        originalCalib1_ = new Pinhole(vCalibration);

        // Rectified images are assumed to be ideal PinHole images (no distortion)
    }
4.如果是广角相机模型

需要在读取相机参数构建相机模型的基础上若是双目模式则要加上重叠区域参数读取以及构建重叠区域。

cpp 复制代码 
else if (cameraModel == "KannalaBrandt8")
    {
        cameraType_ = KannalaBrandt;

        // Read intrinsic parameters
        // 用模板函数读取yaml文件中的参数
        float fx = readParameter<float>(fSettings, "Camera1.fx", found);
        float fy = readParameter<float>(fSettings, "Camera1.fy", found);
        float cx = readParameter<float>(fSettings, "Camera1.cx", found);
        float cy = readParameter<float>(fSettings, "Camera1.cy", found);

        float k0 = readParameter<float>(fSettings, "Camera1.k1", found);
        float k1 = readParameter<float>(fSettings, "Camera1.k2", found);
        float k2 = readParameter<float>(fSettings, "Camera1.k3", found);
        float k3 = readParameter<float>(fSettings, "Camera1.k4", found);

        vCalibration = {fx, fy, cx, cy, k0, k1, k2, k3};

        calibration1_ = new KannalaBrandt8(vCalibration);
        originalCalib1_ = new KannalaBrandt8(vCalibration);
        //双目相机或IMU双目相机时，读取和设置摄像头的图像重叠区域参数。
        if (sensor_ == System::STEREO || sensor_ == System::IMU_STEREO)
        {
            int colBegin = readParameter<int>(fSettings, "Camera1.overlappingBegin", found);
            int colEnd = readParameter<int>(fSettings, "Camera1.overlappingEnd", found);
            vector<int> vOverlapping = {colBegin, colEnd};//摄像头重叠区域。

            static_cast<KannalaBrandt8 *>(calibration1_)->mvLappingArea = vOverlapping;
        }
    }
5.如果没有读取到相机模型

直接输出错误信息。

cpp 复制代码 
else
    {
        cerr << "Error: " << cameraModel << " not known" << endl;
        exit(-1);
    }

完整的代码

cpp 复制代码 
void Settings::readCamera1(cv::FileStorage &fSettings)
{
    bool found;

    // Read camera model
    string cameraModel = readParameter<string>(fSettings, "Camera.type", found);

    vector<float> vCalibration;
    if (cameraModel == "PinHole")
    {
        cameraType_ = PinHole;//设置相机类型

        // Read intrinsic parameters读取内参
        float fx = readParameter<float>(fSettings, "Camera1.fx", found);
        float fy = readParameter<float>(fSettings, "Camera1.fy", found);
        float cx = readParameter<float>(fSettings, "Camera1.cx", found);
        float cy = readParameter<float>(fSettings, "Camera1.cy", found);

        vCalibration = {fx, fy, cx, cy};//储存在校准的变量中

        calibration1_ = new Pinhole(vCalibration);
        originalCalib1_ = new Pinhole(vCalibration);

        // Check if it is a distorted PinHole
        // 检查这个针孔相机的配置文件是否有畸变参数
        readParameter<float>(fSettings, "Camera1.k1", found, false);
        if (found)
        {
            readParameter<float>(fSettings, "Camera1.k3", found, false);
            if (found)
            {
                vPinHoleDistorsion1_.resize(5);//储存相机畸变参数的向量扩容。
                vPinHoleDistorsion1_[4] = readParameter<float>(fSettings, "Camera1.k3", found);
            }
            else
            {
                vPinHoleDistorsion1_.resize(4);
            }
            vPinHoleDistorsion1_[0] = readParameter<float>(fSettings, "Camera1.k1", found);
            vPinHoleDistorsion1_[1] = readParameter<float>(fSettings, "Camera1.k2", found);
            vPinHoleDistorsion1_[2] = readParameter<float>(fSettings, "Camera1.p1", found);
            vPinHoleDistorsion1_[3] = readParameter<float>(fSettings, "Camera1.p2", found);
        }

        // Check if we need to correct distortion from the images
        // 如果是单目相机或者是RGBD相机且有畸变参数则需要去畸变操作。
        if (
            (sensor_ == System::MONOCULAR ||
                sensor_ == System::IMU_MONOCULAR ||
                sensor_ == System::RGBD ||
                sensor_ == System::IMU_RGBD) &&
            vPinHoleDistorsion1_.size() != 0)
        {
            bNeedToUndistort_ = true;
        }
    }
    //如果是校正后的图像则认为没畸变。
    else if (cameraModel == "Rectified")
    {
        cameraType_ = Rectified;

        // Read intrinsic parameters
        float fx = readParameter<float>(fSettings, "Camera1.fx", found);
        float fy = readParameter<float>(fSettings, "Camera1.fy", found);
        float cx = readParameter<float>(fSettings, "Camera1.cx", found);
        float cy = readParameter<float>(fSettings, "Camera1.cy", found);

        vCalibration = {fx, fy, cx, cy};

        calibration1_ = new Pinhole(vCalibration);
        originalCalib1_ = new Pinhole(vCalibration);

        // Rectified images are assumed to be ideal PinHole images (no distortion)
    }
    else if (cameraModel == "KannalaBrandt8")
    {
        cameraType_ = KannalaBrandt;

        // Read intrinsic parameters
        // 用模板函数读取yaml文件中的参数
        float fx = readParameter<float>(fSettings, "Camera1.fx", found);
        float fy = readParameter<float>(fSettings, "Camera1.fy", found);
        float cx = readParameter<float>(fSettings, "Camera1.cx", found);
        float cy = readParameter<float>(fSettings, "Camera1.cy", found);

        float k0 = readParameter<float>(fSettings, "Camera1.k1", found);
        float k1 = readParameter<float>(fSettings, "Camera1.k2", found);
        float k2 = readParameter<float>(fSettings, "Camera1.k3", found);
        float k3 = readParameter<float>(fSettings, "Camera1.k4", found);

        vCalibration = {fx, fy, cx, cy, k0, k1, k2, k3};

        calibration1_ = new KannalaBrandt8(vCalibration);
        originalCalib1_ = new KannalaBrandt8(vCalibration);
        //双目相机或IMU双目相机时，读取和设置摄像头的图像重叠区域参数。
        if (sensor_ == System::STEREO || sensor_ == System::IMU_STEREO)
        {
            int colBegin = readParameter<int>(fSettings, "Camera1.overlappingBegin", found);
            int colEnd = readParameter<int>(fSettings, "Camera1.overlappingEnd", found);
            vector<int> vOverlapping = {colBegin, colEnd};//摄像头重叠区域。

            static_cast<KannalaBrandt8 *>(calibration1_)->mvLappingArea = vOverlapping;
        }
    }
    else
    {
        cerr << "Error: " << cameraModel << " not known" << endl;
        exit(-1);
    }
}

相机2

这时候已经默认是双目模式了，除了要进行相机1的操作，还要额外进行一些操作。

读取基线，计算基线焦距，读取深度。

cpp 复制代码 
// Load stereo extrinsic calibration
    if (cameraType_ == Rectified)
    {
        b_ = readParameter<float>(fSettings, "Stereo.b", found);
        bf_ = b_ * calibration1_->getParameter(0);
    }
    else
    {
        cv::Mat cvTlr = readParameter<cv::Mat>(fSettings, "Stereo.T_c1_c2", found);
        Tlr_ = Converter::toSophus(cvTlr);

        // TODO: also search for Trl and invert if necessary

        b_ = Tlr_.translation().norm();
        bf_ = b_ * calibration1_->getParameter(0);
    }

    thDepth_ = readParameter<float>(fSettings, "Stereo.ThDepth", found);

完整的代码

cpp 复制代码 
void Settings::readCamera2(cv::FileStorage &fSettings)
{
    bool found;
    vector<float> vCalibration;
    if (cameraType_ == PinHole)
    {
        bNeedToRectify_ = true;

        // Read intrinsic parameters
        float fx = readParameter<float>(fSettings, "Camera2.fx", found);
        float fy = readParameter<float>(fSettings, "Camera2.fy", found);
        float cx = readParameter<float>(fSettings, "Camera2.cx", found);
        float cy = readParameter<float>(fSettings, "Camera2.cy", found);

        vCalibration = {fx, fy, cx, cy};

        calibration2_ = new Pinhole(vCalibration);
        originalCalib2_ = new Pinhole(vCalibration);

        // Check if it is a distorted PinHole
        readParameter<float>(fSettings, "Camera2.k1", found, false);
        if (found)
        {
            readParameter<float>(fSettings, "Camera2.k3", found, false);
            if (found)
            {
                vPinHoleDistorsion2_.resize(5);
                vPinHoleDistorsion2_[4] = readParameter<float>(fSettings, "Camera2.k3", found);
            }
            else
            {
                vPinHoleDistorsion2_.resize(4);
            }
            vPinHoleDistorsion2_[0] = readParameter<float>(fSettings, "Camera2.k1", found);
            vPinHoleDistorsion2_[1] = readParameter<float>(fSettings, "Camera2.k2", found);
            vPinHoleDistorsion2_[2] = readParameter<float>(fSettings, "Camera2.p1", found);
            vPinHoleDistorsion2_[3] = readParameter<float>(fSettings, "Camera2.p2", found);
        }
    }
    else if (cameraType_ == KannalaBrandt)
    {
        // Read intrinsic parameters
        float fx = readParameter<float>(fSettings, "Camera2.fx", found);
        float fy = readParameter<float>(fSettings, "Camera2.fy", found);
        float cx = readParameter<float>(fSettings, "Camera2.cx", found);
        float cy = readParameter<float>(fSettings, "Camera2.cy", found);

        float k0 = readParameter<float>(fSettings, "Camera1.k1", found);
        float k1 = readParameter<float>(fSettings, "Camera1.k2", found);
        float k2 = readParameter<float>(fSettings, "Camera1.k3", found);
        float k3 = readParameter<float>(fSettings, "Camera1.k4", found);

        vCalibration = {fx, fy, cx, cy, k0, k1, k2, k3};

        calibration2_ = new KannalaBrandt8(vCalibration);
        originalCalib2_ = new KannalaBrandt8(vCalibration);

        int colBegin = readParameter<int>(fSettings, "Camera2.overlappingBegin", found);
        int colEnd = readParameter<int>(fSettings, "Camera2.overlappingEnd", found);
        vector<int> vOverlapping = {colBegin, colEnd};

        static_cast<KannalaBrandt8 *>(calibration2_)->mvLappingArea = vOverlapping;
    }

    // Load stereo extrinsic calibration
    if (cameraType_ == Rectified)
    {
        b_ = readParameter<float>(fSettings, "Stereo.b", found);
        bf_ = b_ * calibration1_->getParameter(0);
    }
    else
    {
        cv::Mat cvTlr = readParameter<cv::Mat>(fSettings, "Stereo.T_c1_c2", found);
        Tlr_ = Converter::toSophus(cvTlr);

        // TODO: also search for Trl and invert if necessary

        b_ = Tlr_.translation().norm();
        bf_ = b_ * calibration1_->getParameter(0);
    }

    thDepth_ = readParameter<float>(fSettings, "Stereo.ThDepth", found);
}

结束语

以上就是我学习到的内容，如果对您有帮助请多多支持我，如果哪里有问题欢迎大家在评论区积极讨论，我看到会及时回复。

相关推荐
羑悻的小杀马特1 小时前
OpenCV 引擎：驱动实时应用开发的科技狂飙
人工智能·科技·opencv·计算机视觉
蹦蹦跳跳真可爱5892 小时前
Python----计算机视觉处理（Opencv:道路检测之提取车道线）
python·opencv·计算机视觉
SKYDROID云卓小助手2 小时前
三轴云台之相机技术篇
运维·服务器·网络·数码相机·音视频
guanshiyishi4 小时前
ABeam 德硕 | 中国汽车市场(2)——新能源车的崛起与中国汽车市场机遇与挑战
人工智能
极客天成ScaleFlash4 小时前
极客天成NVFile：无缓存直击存储性能天花板，重新定义AI时代并行存储新范式
人工智能·缓存
澳鹏Appen5 小时前
AI安全：构建负责任且可靠的系统
人工智能·安全
蹦蹦跳跳真可爱5896 小时前
Python----机器学习（KNN：使用数学方法实现KNN）
人工智能·python·机器学习
视界宝藏库6 小时前
多元 AI 配音软件，打造独特音频体验
人工智能
xinxiyinhe7 小时前
GitHub上英语学习工具的精选分类汇总
人工智能·deepseek·学习英语精选
zhuyixiangyyds7 小时前
day21和day22学习Pandas库
笔记·学习·pandas
热门推荐
01我决定放弃搞 Java 了02DeepSeek各版本说明与优缺点分析03如何在WPS和Word/Excel中直接使用DeepSeek功能04RAG 实践- Ollama+RagFlow 部署本地知识库05本地化部署AI知识库：基于Ollama+DeepSeek+AnythingLLM保姆级教程06从零安装 LLaMA-Factory 微调 Qwen 大模型成功及所有的坑07如何本地部署AI智能体平台，带你手搓一个AI Agent08苍穹外卖面试总结09DeepSeek RAGFlow构建本地知识库系统10【芯片封测学习专栏 -- D2D 和 C2C 之间的区别】