先要拿大名鼎鼎的ORB-SLAM系統框圖鎮樓，看著這張圖能夠完美的串起來整個流程。

ORB-SLAM分三個執行緒，分別是Tracking、LocalMapping和LoopClosing。

（1）Tracking：在主執行緒上，輸入視訊流，輸出相機位姿並跟蹤區域性地圖。提取ORB特徵子，根據上一幀進行位姿估計或全域性重定位，然後跟蹤區域性地圖優化位姿，確定新的關鍵幀。

（2）LocalMapping：維護優化區域性地圖。輸入關鍵幀，輸出修正的關鍵幀和地圖點。插入關鍵幀，生成新的地圖點，使用BA優化，去除冗餘的關鍵幀

（3）LoopClosing：閉環檢測和閉環矯正。BOW檢測，再用Sim3計算相似變換；閉環融合與優化Essential Graph的圖優化。

#ifndef SYSTEM_H
#define SYSTEM_H

#include<string>
#include<thread>
#include<opencv2/core/core.hpp>

#include "Tracking.h"
#include "FrameDrawer.h"
#include "MapDrawer.h"
#include "Map.h"
#include "LocalMapping.h"
#include "LoopClosing.h"
#include "KeyFrameDatabase.h"
#include "ORBVocabulary.h"
#include "Viewer.h"

namespace ORB_SLAM2
{

class Viewer;
class FrameDrawer;
class Map;
class Tracking;
class LocalMapping;
class LoopClosing;

class System
{
public:
    // 三種感測器選擇
    enum eSensor{
        MONOCULAR=0,
        STEREO=1,
        RGBD=2
    };
	//成員函式
public:
    // 宣告
    System(const string &strVocFile, const string &strSettingsFile, const eSensor sensor, const bool bUseViewer = true);

    // Tracking函式:輸出相機位姿
    cv::Mat TrackStereo(const cv::Mat &imLeft, const cv::Mat &imRight, const double &timestamp);
    cv::Mat TrackRGBD(const cv::Mat &im, const cv::Mat &depthmap, const double &timestamp);
    cv::Mat TrackMonocular(const cv::Mat &im, const double &timestamp);

    // 啟用定位模組；停止Local Mapping，只相機追蹤？？？
    void ActivateLocalizationMode();
    // 失效定位模組；恢復Local Mapping,再次執行SLAM？？？
    void DeactivateLocalizationMode();
    // 清空地圖，重啟系統
    void Reset();
    // 儲存軌跡之前執行
    void Shutdown();
    // 儲存相機軌跡 Only stereo and RGB-D. 
    // Call first Shutdown()
    // See format details at: http://vision.in.tum.de/data/datasets/rgbd-dataset
    void SaveTrajectoryTUM(const string &filename);
	// See format details at: http://www.cvlibs.net/datasets/kitti/eval_odometry.php
	void SaveTrajectoryKITTI(const string &filename);
    // 儲存關鍵幀位姿 
    // Call first Shutdown()
    // See format details at: http://vision.in.tum.de/data/datasets/rgbd-dataset
    void SaveKeyFrameTrajectoryTUM(const string &filename);

	//成員變數
private:
    // Input sensor
    eSensor mSensor;

    // ORB詞彙表用於場景識別和特徵匹配
    ORBVocabulary* mpVocabulary;

    // 關鍵幀資料庫用於位置識別 (重定位和迴環檢測).
    KeyFrameDatabase* mpKeyFrameDatabase;

    // 儲存關鍵幀和地圖特徵子
    Map* mpMap;

    // Tracker. 接受幀計算相機位姿
    // 決定何時插入新的關鍵幀，建立新的地圖特徵子
    // 重定位
    Tracking* mpTracker;
    // Local Mapper.管理本地地圖，執行BA
    LocalMapping* mpLocalMapper;
    // Loop Closer. 搜尋每個新的關鍵幀的迴圈
    LoopClosing* mpLoopCloser;
    // Pangolin.繪製地圖和當前相機位姿
    Viewer* mpViewer;
	// 畫圖用的
    FrameDrawer* mpFrameDrawer;
    MapDrawer* mpMapDrawer;

    // 3個執行緒: Local Mapping, Loop Closing, Viewer.
    // Tracking執行緒在System主程式執行緒中
    std::thread* mptLocalMapping;
    std::thread* mptLoopClosing;
    std::thread* mptViewer;

    // 重置Flag？？？
    std::mutex mMutexReset;
    bool mbReset;

    //更改模型flags
    std::mutex mMutexMode;
    bool mbActivateLocalizationMode;
    bool mbDeactivateLocalizationMode;
};

}// namespace ORB_SLAM

#endif // SYSTEM_H
 

#include "System.h"
#include "Converter.h"
#include <thread>
#include <pangolin/pangolin.h>
#include <iostream>     // std::cout, std::fixed
#include <iomanip>		// std::setprecision
bool has_suffix(const std::string &str, const std::string &suffix) {
  std::size_t index = str.find(suffix, str.size() - suffix.size());
  return (index != std::string::npos);
}

namespace ORB_SLAM2
{
	//列表初始化建構函式：詞袋檔案、引數檔案、感測器型別、是否用Viewer
System::System(const string &strVocFile, const string &strSettingsFile, const eSensor sensor,
               const bool bUseViewer):mSensor(sensor),mbReset(false),mbActivateLocalizationMode(false),
               mbDeactivateLocalizationMode(false)
{
    // Output welcome message
    cout << endl <<
    "ORB-SLAM2 Copyright (C) 2014-2016 Raul Mur-Artal, University of Zaragoza." << endl <<
    "This program comes with ABSOLUTELY NO WARRANTY;" << endl  <<
    "This is free software, and you are welcome to redistribute it" << endl <<
    "under certain conditions. See LICENSE.txt." << endl << endl;

    cout << "Input sensor was set to: ";

    if(mSensor==MONOCULAR)
        cout << "Monocular" << endl;
    else if(mSensor==STEREO)
        cout << "Stereo" << endl;
    else if(mSensor==RGBD)
        cout << "RGB-D" << endl;

    //1.讀取引數檔案，內參、幀率、基線、深度, XXX.yaml
    cv::FileStorage fsSettings(strSettingsFile.c_str(), cv::FileStorage::READ);
    if(!fsSettings.isOpened())
    {
       cerr << "Failed to open settings file at: " << strSettingsFile << endl;
       exit(-1);
    }

    //2.下載ORB詞袋  .txt
    cout << endl << "Loading ORB Vocabulary. This could take a while..." << endl;
    mpVocabulary = new ORBVocabulary();//這個詞袋類在哪裡定義的？？？
    bool bVocLoad = false; // 基於副檔名載入
    if (has_suffix(strVocFile, ".txt"))
	  bVocLoad = mpVocabulary->loadFromTextFile(strVocFile);
	else if(has_suffix(strVocFile, ".bin"))
	  bVocLoad = mpVocabulary->loadFromBinaryFile(strVocFile);
	else
	  bVocLoad = false;
    if(!bVocLoad)
    {
        cerr << "Wrong path to vocabulary. " << endl;
        cerr << "Failed to open at: " << strVocFile << endl;
        exit(-1);
    }
    cout << "Vocabulary loaded!" << endl << endl;

    //3.建立關鍵幀資料庫
    mpKeyFrameDatabase = new KeyFrameDatabase(*mpVocabulary);

    //4.建立地圖
    mpMap = new Map();

    //Create Drawers.
    mpFrameDrawer = new FrameDrawer(mpMap);
    mpMapDrawer = new MapDrawer(mpMap, strSettingsFile);

    //5.1初始化 Tracking 
    //(it will live in the main thread of execution, the one that called this constructor)
    mpTracker = new Tracking(this, mpVocabulary, mpFrameDrawer, mpMapDrawer,
                             mpMap, mpKeyFrameDatabase, strSettingsFile, mSensor);

    //5.2初始化併發布 Local Mapping 執行緒
    mpLocalMapper = new LocalMapping(mpMap, mSensor==MONOCULAR);
    mptLocalMapping = new thread(&ORB_SLAM2::LocalMapping::Run,mpLocalMapper);

    //5.3初始化併發布 Loop Closing 執行緒
    mpLoopCloser = new LoopClosing(mpMap, mpKeyFrameDatabase, mpVocabulary, mSensor!=MONOCULAR);
    mptLoopClosing = new thread(&ORB_SLAM2::LoopClosing::Run, mpLoopCloser);

    //5.4初始化併發布 Viewer 執行緒
    mpViewer = new Viewer(this, mpFrameDrawer,mpMapDrawer,mpTracker,strSettingsFile);
    if(bUseViewer)
        mptViewer = new thread(&Viewer::Run, mpViewer);

    mpTracker->SetViewer(mpViewer);

    //5.5執行緒之間相互設定指標？？？
    mpTracker->SetLocalMapper(mpLocalMapper);
    mpTracker->SetLoopClosing(mpLoopCloser);

    mpLocalMapper->SetTracker(mpTracker);
    mpLocalMapper->SetLoopCloser(mpLoopCloser);

    mpLoopCloser->SetTracker(mpTracker);
    mpLoopCloser->SetLocalMapper(mpLocalMapper);
}

cv::Mat System::TrackStereo(const cv::Mat &imLeft, const cv::Mat &imRight, const double &timestamp)
{
    if(mSensor!=STEREO)
    {
        cerr << "ERROR: you called TrackStereo but input sensor was not set to STEREO." << endl;
        exit(-1);
    }   

    // Check mode change
    {
        unique_lock<mutex> lock(mMutexMode);
        if(mbActivateLocalizationMode)
        {
            mpLocalMapper->RequestStop();

            // Wait until Local Mapping has effectively stopped
            while(!mpLocalMapper->isStopped())
            {
                //usleep(1000);
				std::this_thread::sleep_for(std::chrono::milliseconds(1));
            }

            mpTracker->InformOnlyTracking(true);// 定位時，只跟蹤
            mbActivateLocalizationMode = false;
        }
        if(mbDeactivateLocalizationMode)
        {
            mpTracker->InformOnlyTracking(false);
            mpLocalMapper->Release();
            mbDeactivateLocalizationMode = false;
        }
    }

    // Check reset
    {
    unique_lock<mutex> lock(mMutexReset);
    if(mbReset)
    {
        mpTracker->Reset();
        mbReset = false;
    }
    }

    return mpTracker->GrabImageStereo(imLeft,imRight,timestamp);
}

cv::Mat System::TrackRGBD(const cv::Mat &im, const cv::Mat &depthmap, const double &timestamp)
{
    if(mSensor!=RGBD)
    {
        cerr << "ERROR: you called TrackRGBD but input sensor was not set to RGBD." << endl;
        exit(-1);
    }    

    // Check mode change
    {
        unique_lock<mutex> lock(mMutexMode);
        if(mbActivateLocalizationMode)
        {
            mpLocalMapper->RequestStop();

            // Wait until Local Mapping has effectively stopped
            while(!mpLocalMapper->isStopped())
            {
                //usleep(1000);
				std::this_thread::sleep_for(std::chrono::milliseconds(1));
			}

            mpTracker->InformOnlyTracking(true);// 定位時，只跟蹤
            mbActivateLocalizationMode = false;
        }
        if(mbDeactivateLocalizationMode)
        {
            mpTracker->InformOnlyTracking(false);
            mpLocalMapper->Release();
            mbDeactivateLocalizationMode = false;
        }
    }

    // Check reset
    {
    unique_lock<mutex> lock(mMutexReset);
    if(mbReset)
    {
        mpTracker->Reset();
        mbReset = false;
    }
    }

    return mpTracker->GrabImageRGBD(im,depthmap,timestamp);
}
//定義跟蹤單目函式
cv::Mat System::TrackMonocular(const cv::Mat &im, const double &timestamp)
{
    if(mSensor!=MONOCULAR)
    {
        cerr << "ERROR: you called TrackMonocular but input sensor was not set to Monocular." << endl;
        exit(-1);
    }

    // 檢查模型
    {
        unique_lock<mutex> lock(mMutexMode);
		//如果啟用定位模組，休眠1000ms直到停止建圖
        if(mbActivateLocalizationMode)
        {
            mpLocalMapper->RequestStop();
            // Wait until Local Mapping has effectively stopped
            while(!mpLocalMapper->isStopped())
            {
                //usleep(1000);
                std::this_thread::sleep_for(std::chrono::milliseconds(1));
            }

            mpTracker->InformOnlyTracking(true);// 定位時，只跟蹤
            mbActivateLocalizationMode = false;// 防止重複執行
        }
		//如果定位模組失效，重啟建圖
        if(mbDeactivateLocalizationMode)
        {
            mpTracker->InformOnlyTracking(false);
            mpLocalMapper->Release();
            mbDeactivateLocalizationMode = false;// 防止重複執行
        }
    }

    // 檢查重啟
    {
        unique_lock<mutex> lock(mMutexReset);
        if(mbReset)
        {
            mpTracker->Reset();
            mbReset = false;
        }
    }
	//這個Grab函式哪裡來的？？？
    return mpTracker->GrabImageMonocular(im,timestamp);
}
//啟用定位模組
void System::ActivateLocalizationMode()
{
    unique_lock<mutex> lock(mMutexMode);
    mbActivateLocalizationMode = true;
}
//失效定位模組
void System::DeactivateLocalizationMode()
{
	//加鎖
    unique_lock<mutex> lock(mMutexMode);
    mbDeactivateLocalizationMode = true;
}
//重置系統
void System::Reset()
{
    unique_lock<mutex> lock(mMutexReset);
    mbReset = true;
}
//關閉整個系統
void System::Shutdown()
{
    mpLocalMapper->RequestFinish();
    mpLoopCloser->RequestFinish();
    if(mpViewer)
    {
        mpViewer->RequestFinish();
        while(!mpViewer->isFinished())
            std::this_thread::sleep_for(std::chrono::milliseconds(5));
    }

    // Wait until all thread have effectively stopped
    while(!mpLocalMapper->isFinished() || !mpLoopCloser->isFinished() || mpLoopCloser->isRunningGBA())
    {
        //usleep(5000);
		std::this_thread::sleep_for(std::chrono::milliseconds(5));
	}

    if(mpViewer)
        pangolin::BindToContext("ORB-SLAM2: Map Viewer");
}
//儲存軌跡
void System::SaveTrajectoryTUM(const string &filename)
{
    cout << endl << "Saving camera trajectory to " << filename << " ..." << endl;
    if(mSensor==MONOCULAR)
    {
        cerr << "ERROR: SaveTrajectoryTUM cannot be used for monocular." << endl;
        return;
    }
	//得到所有關鍵幀，重新排序第一幀位於原點
    vector<KeyFrame*> vpKFs = mpMap->GetAllKeyFrames();
    sort(vpKFs.begin(),vpKFs.end(),KeyFrame::lId);
    cv::Mat Two = vpKFs[0]->GetPoseInverse();

    ofstream f;
    f.open(filename.c_str());
    f << fixed;

    // 先獲得關鍵幀，當前幀相對於關鍵幀
	//每一幀都有引用關鍵幀、時間戳和標誌
    list<ORB_SLAM2::KeyFrame*>::iterator lRit = mpTracker->mlpReferences.begin();
    list<double>::iterator lT = mpTracker->mlFrameTimes.begin();
    list<bool>::iterator lbL = mpTracker->mlbLost.begin();
    for(list<cv::Mat>::iterator lit=mpTracker->mlRelativeFramePoses.begin(),
        lend=mpTracker->mlRelativeFramePoses.end();lit!=lend;lit++, lRit++, lT++, lbL++)
    {
        if(*lbL)
            continue;

        KeyFrame* pKF = *lRit;

        cv::Mat Trw = cv::Mat::eye(4,4,CV_32F);

        // 如果參考關鍵幀被剔除，遍歷生成樹獲得合適關鍵幀
        while(pKF->isBad())
        {
            Trw = Trw*pKF->mTcp;
            pKF = pKF->GetParent();
        }

        Trw = Trw*pKF->GetPose()*Two;

        cv::Mat Tcw = (*lit)*Trw;
        cv::Mat Rwc = Tcw.rowRange(0,3).colRange(0,3).t();
        cv::Mat twc = -Rwc*Tcw.rowRange(0,3).col(3);

        vector<float> q = Converter::toQuaternion(Rwc);

        f << setprecision(6) << *lT << " " <<  setprecision(9) << twc.at<float>(0) << " " << twc.at<float>(1) << " " << twc.at<float>(2) << " " << q[0] << " " << q[1] << " " << q[2] << " " << q[3] << endl;
    }
    f.close();
    cout << endl << "trajectory saved!" << endl;
}


void System::SaveKeyFrameTrajectoryTUM(const string &filename)
{
    cout << endl << "Saving keyframe trajectory to " << filename << " ..." << endl;

    vector<KeyFrame*> vpKFs = mpMap->GetAllKeyFrames();
    sort(vpKFs.begin(),vpKFs.end(),KeyFrame::lId);

    // Transform all keyframes so that the first keyframe is at the origin.
    // After a loop closure the first keyframe might not be at the origin.
    //cv::Mat Two = vpKFs[0]->GetPoseInverse();

    ofstream f;
    f.open(filename.c_str());
    f << fixed;

    for(size_t i=0; i<vpKFs.size(); i++)
    {
        KeyFrame* pKF = vpKFs[i];

       // pKF->SetPose(pKF->GetPose()*Two);

        if(pKF->isBad())
            continue;

        cv::Mat R = pKF->GetRotation().t();
        vector<float> q = Converter::toQuaternion(R);
        cv::Mat t = pKF->GetCameraCenter();
        f << setprecision(6) << pKF->mTimeStamp << setprecision(7) << " " << t.at<float>(0) << " " << t.at<float>(1) << " " << t.at<float>(2)
          << " " << q[0] << " " << q[1] << " " << q[2] << " " << q[3] << endl;

    }

    f.close();
    cout << endl << "trajectory saved!" << endl;
}

void System::SaveTrajectoryKITTI(const string &filename)
{
    cout << endl << "Saving camera trajectory to " << filename << " ..." << endl;
    if(mSensor==MONOCULAR)
    {
        cerr << "ERROR: SaveTrajectoryKITTI cannot be used for monocular." << endl;
        return;
    }

    vector<KeyFrame*> vpKFs = mpMap->GetAllKeyFrames();
    sort(vpKFs.begin(),vpKFs.end(),KeyFrame::lId);

    // Transform all keyframes so that the first keyframe is at the origin.
    // After a loop closure the first keyframe might not be at the origin.
    cv::Mat Two = vpKFs[0]->GetPoseInverse();

    ofstream f;
    f.open(filename.c_str());
    f << fixed;

    // Frame pose is stored relative to its reference keyframe (which is optimized by BA and pose graph).
    // We need to get first the keyframe pose and then concatenate the relative transformation.
    // Frames not localized (tracking failure) are not saved.

    // For each frame we have a reference keyframe (lRit), the timestamp (lT) and a flag
    // which is true when tracking failed (lbL).
    list<ORB_SLAM2::KeyFrame*>::iterator lRit = mpTracker->mlpReferences.begin();
    list<double>::iterator lT = mpTracker->mlFrameTimes.begin();
    for(list<cv::Mat>::iterator lit=mpTracker->mlRelativeFramePoses.begin(), lend=mpTracker->mlRelativeFramePoses.end();lit!=lend;lit++, lRit++, lT++)
    {
        ORB_SLAM2::KeyFrame* pKF = *lRit;

        cv::Mat Trw = cv::Mat::eye(4,4,CV_32F);

        while(pKF->isBad())
        {
          //  cout << "bad parent" << endl;
            Trw = Trw*pKF->mTcp;
            pKF = pKF->GetParent();
        }

        Trw = Trw*pKF->GetPose()*Two;

        cv::Mat Tcw = (*lit)*Trw;
        cv::Mat Rwc = Tcw.rowRange(0,3).colRange(0,3).t();
        cv::Mat twc = -Rwc*Tcw.rowRange(0,3).col(3);

        f << setprecision(9) << Rwc.at<float>(0,0) << " " << Rwc.at<float>(0,1)  << " " << Rwc.at<float>(0,2) << " "  << twc.at<float>(0) << " " <<
             Rwc.at<float>(1,0) << " " << Rwc.at<float>(1,1)  << " " << Rwc.at<float>(1,2) << " "  << twc.at<float>(1) << " " <<
             Rwc.at<float>(2,0) << " " << Rwc.at<float>(2,1)  << " " << Rwc.at<float>(2,2) << " "  << twc.at<float>(2) << endl;
    }
    f.close();
    cout << endl << "trajectory saved!" << endl;
}

} //namespace ORB_SLAM
 

總結：

我們可以看到在這個物件的例項化過程中，我們建立了以下物件： 
（1）建立了ORB詞袋的物件 
（2）建立了關鍵幀的資料庫 
（3）建立地圖物件 
（4）建立兩個顯示視窗 
（5）初始化Tracking物件 
（6）初始化Local Mapping物件，並開啟執行緒執行 
（7）初始化Loop Closing物件，並開啟執行緒執行 
（8）初始化視窗，開啟執行緒顯示影象和地圖點