ORBSLAM2源码学习（8） LocalMapping类

最近有点事博客更新断了几天，接下来争取一次性更完。LocalMapping这个类维护了一个局部地图，对关键帧进行处理、地图点检查剔除、生成新地图点。还是先上代码再总结。

#include "LocalMapping.h"
#include "LoopClosing.h"
#include "ORBmatcher.h"
#include "Optimizer.h"

#include<mutex>

namespace ORB_SLAM2
{

LocalMapping::LocalMapping(Map *pMap, const float bMonocular):
    mbMonocular(bMonocular), mbResetRequested(false), mbFinishRequested(false), mbFinished(true), mpMap(pMap),
    mbAbortBA(false), mbStopped(false), mbStopRequested(false), mbNotStop(false), mbAcceptKeyFrames(true)
{
}

void LocalMapping::SetLoopCloser(LoopClosing* pLoopCloser)
{
    mpLoopCloser = pLoopCloser;
}

void LocalMapping::SetTracker(Tracking *pTracker)
{
    mpTracker=pTracker;
}

void LocalMapping::Run()
{

    mbFinished = false;

    while(1)
    {
        // Tracking will see that Local Mapping is busy
        // LocalMapping处于繁忙状态，处理的关键帧是Tracking线程发过来的
        // 在LocalMapping线程处理完关键帧之前Tracking线程不能发送太快
        SetAcceptKeyFrames(false);

        // Check if there are keyframes in the queue
        // 等待处理的关键帧不为空
        if(CheckNewKeyFrames())
        {
            // BoW conversion and insertion in Map
            // 计算关键帧特征点的BoW，将关键帧插入地图
            ProcessNewKeyFrame();

            // Check recent MapPoints
            // 剔除ProcessNewKeyFrame函数中引入的不合格MapPoints
            MapPointCulling();

            // Triangulate new MapPoints
            // 相机运动过程中与相邻关键帧通过三角化恢复出MapPoints
            CreateNewMapPoints();

            // 已经处理完队列中的最后的一个关键帧
            if(!CheckNewKeyFrames())
            {
                // Find more matches in neighbor keyframes and fuse point duplications
                // 检查并融合当前关键帧与相邻帧重复的MapPoints
                SearchInNeighbors();
            }

            mbAbortBA = false;

            // 已经处理完队列中的最后的一个关键帧，并且闭环检测没有请求停止LocalMapping，则进行一个local BA
            if(!CheckNewKeyFrames() && !stopRequested())
            {
                if(mpMap->KeyFramesInMap()>2)
                    Optimizer::LocalBundleAdjustment(mpCurrentKeyFrame,&mbAbortBA, mpMap);

                // Check redundant local Keyframes
                // 检测并剔除当前帧相邻的关键帧中冗余的关键帧
                // 标准：该关键帧的90%的MapPoints可以被其它关键帧观测到
                // 在Tracking中插入关键帧的条件比较松，交给LocalMapping线程的关键帧会比较多
                // 在这里再剔除冗余的关键帧
                KeyFrameCulling();
            }

            // 将当前帧加入到闭环检测队列中
            mpLoopCloser->InsertKeyFrame(mpCurrentKeyFrame);
        }
        else if(Stop())
        {
            // Safe area to stop
            while(isStopped() && !CheckFinish())
            {
                // usleep(3000);
                std::this_thread::sleep_for(std::chrono::milliseconds(3));
            }
            if(CheckFinish())
                break;
        }

        ResetIfRequested();

        // Tracking will see that Local Mapping is not busy
        SetAcceptKeyFrames(true);

        if(CheckFinish())
            break;

        //usleep(3000);
        std::this_thread::sleep_for(std::chrono::milliseconds(3));

    }

    SetFinish();
}

// 将关键帧插入到地图中，以便将来进行局部地图优化
void LocalMapping::InsertKeyFrame(KeyFrame *pKF)
{
    unique_lock<mutex> lock(mMutexNewKFs);
    // 将关键帧插入到列表中
    mlNewKeyFrames.push_back(pKF);
    mbAbortBA=true;
}

// 看关键帧是否为空
bool LocalMapping::CheckNewKeyFrames()
{
    unique_lock<mutex> lock(mMutexNewKFs);
    return(!mlNewKeyFrames.empty());
}

// 处理列表中的关键帧
// 关联当前关键帧至MapPoints，并更新MapPoints的平均观测方向和观测距离范围
void LocalMapping::ProcessNewKeyFrame()
{
    // 从队列中取出一帧关键帧,最前面那一帧
    {
        unique_lock<mutex> lock(mMutexNewKFs);
        mpCurrentKeyFrame = mlNewKeyFrames.front();
        mlNewKeyFrames.pop_front();
    }

    // Compute Bags of Words structures
    // 计算该关键帧特征点的Bow
    mpCurrentKeyFrame->ComputeBoW();

    // Associate MapPoints to the new keyframe and update normal and descriptor
    // 跟踪局部地图过程中新匹配上的MapPoints和当前关键帧绑定
    // 在TrackLocalMap函数中将局部地图中的MapPoints与当前帧进行了匹配，但没有与当前帧进行关联
    const vector<MapPoint*> vpMapPointMatches = mpCurrentKeyFrame->GetMapPointMatches();

    for(size_t i=0; i<vpMapPointMatches.size(); i++)
    {
        MapPoint* pMP = vpMapPointMatches[i];
        if(pMP)
        {
            if(!pMP->isBad())
            {
                // 非当前帧生成的地图点
				// 为当前帧在tracking过程跟踪到的地图点添加属性
                if(!pMP->IsInKeyFrame(mpCurrentKeyFrame))
                {
                    // 添加观测
                    pMP->AddObservation(mpCurrentKeyFrame, i);
                    // 获得该点的平均观测方向和观测距离范围
                    pMP->UpdateNormalAndDepth();
                    // 加入关键帧后，更新3d点的最佳描述子
                    pMP->ComputeDistinctiveDescriptors();
                }
                else // this can only happen for new stereo points inserted by the Tracking
                {
                    mlpRecentAddedMapPoints.push_back(pMP);
                }
            }
        }
    }    

    // Update links in the Covisibility Graph
    // 更新关键帧间的连接关系，Covisibility图和Essential图
    mpCurrentKeyFrame->UpdateConnections();

    // Insert Keyframe in Map
    // 将该关键帧插入到地图中
    mpMap->AddKeyFrame(mpCurrentKeyFrame);
}

// 剔除ProcessNewKeyFrame和CreateNewMapPoints函数中引入的质量不好的MapPoints
void LocalMapping::MapPointCulling()
{
    // Check Recent Added MapPoints
    list<MapPoint*>::iterator lit = mlpRecentAddedMapPoints.begin();
    const unsigned long int nCurrentKFid = mpCurrentKeyFrame->mnId;

    int nThObs;
    if(mbMonocular)
        nThObs = 2;
    else
        nThObs = 3;
    const int cnThObs = nThObs;
	
	// 遍历等待检查的MapPoints
    while(lit!=mlpRecentAddedMapPoints.end())
    {
        MapPoint* pMP = *lit;
        if(pMP->isBad())
        {
            // 已经是坏点的MapPoints直接从链表中删除
            lit = mlpRecentAddedMapPoints.erase(lit);
        }
        else if(pMP->GetFoundRatio()<0.25f)
        {
            // 跟踪到该MapPoint的Frame数相比预计可观测到该MapPoint的Frame数的比例小于25%
            pMP->SetBadFlag();
            lit = mlpRecentAddedMapPoints.erase(lit);
        }
        else if(((int)nCurrentKFid-(int)pMP->mnFirstKFid)>=2 && pMP->Observations()<=cnThObs)
        {
            // 从该点建立开始，到现在已经过了不少于2个关键帧
            // 但是观测到该点的关键帧数却不超过cnThObs帧，则该点不合格
            pMP->SetBadFlag();
            lit = mlpRecentAddedMapPoints.erase(lit);
        }
        else if(((int)nCurrentKFid-(int)pMP->mnFirstKFid)>=3)
            // 从建立该点开始，已经过了3个关键帧而没有被剔除，则认为是质量高的点
            // 仅从队列中删除，放弃对该点的检测
            lit = mlpRecentAddedMapPoints.erase(lit);
        else
            lit++;
    }
}

// 产生新的地图点
void LocalMapping::CreateNewMapPoints()
{
    // Retrieve neighbor keyframes in covisibility graph
    int nn = 10;
    if(mbMonocular)
        nn=20;
    // 在当前关键帧的共视关键帧中找到共视程度最高的nn帧相邻帧
    const vector<KeyFrame*> vpNeighKFs = mpCurrentKeyFrame->GetBestCovisibilityKeyFrames(nn);

    ORBmatcher matcher(0.6,false);

    cv::Mat Rcw1 = mpCurrentKeyFrame->GetRotation();
    cv::Mat Rwc1 = Rcw1.t();
    cv::Mat tcw1 = mpCurrentKeyFrame->GetTranslation();
    cv::Mat Tcw1(3,4,CV_32F);
    Rcw1.copyTo(Tcw1.colRange(0,3));
    tcw1.copyTo(Tcw1.col(3));

    cv::Mat Ow1 = mpCurrentKeyFrame->GetCameraCenter();

    const float &fx1 = mpCurrentKeyFrame->fx;
    const float &fy1 = mpCurrentKeyFrame->fy;
    const float &cx1 = mpCurrentKeyFrame->cx;
    const float &cy1 = mpCurrentKeyFrame->cy;
    const float &invfx1 = mpCurrentKeyFrame->invfx;
    const float &invfy1 = mpCurrentKeyFrame->invfy;

    const float ratioFactor = 1.5f*mpCurrentKeyFrame->mfScaleFactor;

    int nnew=0;

    // Search matches with epipolar restriction and triangulate
    // 遍历相邻关键帧
    for(size_t i=0; i<vpNeighKFs.size(); i++)
    {
        if(i>0 && CheckNewKeyFrames())
            return;

        KeyFrame* pKF2 = vpNeighKFs[i];

        // Check first that baseline is not too short
        cv::Mat Ow2 = pKF2->GetCameraCenter();
        // 两个关键帧间的相机位移
        cv::Mat vBaseline = Ow2-Ow1;
        const float baseline = cv::norm(vBaseline);

        // 判断相机运动的基线是不是足够长
        if(!mbMonocular)
        {
            // 如果是立体相机，关键帧间距太小时不生成3D点
            if(baseline<pKF2->mb)
            continue;
        }
        else
        {
            // 邻接关键帧的场景深度中值
            const float medianDepthKF2 = pKF2->ComputeSceneMedianDepth(2);
            // baseline与景深的比例
            const float ratioBaselineDepth = baseline/medianDepthKF2;
            // 如果特别远，那么不考虑当前邻接的关键帧，不生成3D点
            if(ratioBaselineDepth<0.01)
                continue;
        }

        // Compute Fundamental Matrix
        // 根据两个关键帧的位姿计算它们之间的基本矩阵
        cv::Mat F12 = ComputeF12(mpCurrentKeyFrame,pKF2);

        // Search matches that fullfil epipolar constraint
        // 通过极线约束限制匹配时的搜索范围，进行特征点匹配
        vector<pair<size_t,size_t> > vMatchedIndices;
        matcher.SearchForTriangulation(mpCurrentKeyFrame,pKF2,F12,vMatchedIndices,false);

        cv::Mat Rcw2 = pKF2->GetRotation();
        cv::Mat Rwc2 = Rcw2.t();
        cv::Mat tcw2 = pKF2->GetTranslation();
        cv::Mat Tcw2(3,4,CV_32F);
        Rcw2.copyTo(Tcw2.colRange(0,3));
        tcw2.copyTo(Tcw2.col(3));

        const float &fx2 = pKF2->fx;
        const float &fy2 = pKF2->fy;
        const float &cx2 = pKF2->cx;
        const float &cy2 = pKF2->cy;
        const float &invfx2 = pKF2->invfx;
        const float &invfy2 = pKF2->invfy;

        // Triangulate each match
        // 对每对匹配通过三角化生成3D点
        const int nmatches = vMatchedIndices.size();
        for(int ikp=0; ikp<nmatches; ikp++)
        {
            // 当前关键帧中的索引
            const int &idx1 = vMatchedIndices[ikp].first;
            
            // 邻接关键帧中的索引
            const int &idx2 = vMatchedIndices[ikp].second;

            const cv::KeyPoint &kp1 = mpCurrentKeyFrame->mvKeysUn[idx1];
            const float kp1_ur=mpCurrentKeyFrame->mvuRight[idx1];
            bool bStereo1 = kp1_ur>=0;

            const cv::KeyPoint &kp2 = pKF2->mvKeysUn[idx2];
            const float kp2_ur = pKF2->mvuRight[idx2];
            bool bStereo2 = kp2_ur>=0;

            // Check parallax between rays
            // 利用匹配点反投影得到视差角
            cv::Mat xn1 = (cv::Mat_<float>(3,1) << (kp1.pt.x-cx1)*invfx1, (kp1.pt.y-cy1)*invfy1, 1.0);
            cv::Mat xn2 = (cv::Mat_<float>(3,1) << (kp2.pt.x-cx2)*invfx2, (kp2.pt.y-cy2)*invfy2, 1.0);

            // 由相机坐标系转到世界坐标系，得到视差角
            cv::Mat ray1 = Rwc1*xn1;
            cv::Mat ray2 = Rwc2*xn2;
            const float cosParallaxRays = ray1.dot(ray2)/(cv::norm(ray1)*cv::norm(ray2));

            float cosParallaxStereo = cosParallaxRays+1;
            float cosParallaxStereo1 = cosParallaxStereo;
            float cosParallaxStereo2 = cosParallaxStereo;

            // 对于双目，利用双目得到视差角
            if(bStereo1)
                cosParallaxStereo1 = cos(2*atan2(mpCurrentKeyFrame->mb/2,mpCurrentKeyFrame->mvDepth[idx1]));
            else if(bStereo2)
                cosParallaxStereo2 = cos(2*atan2(pKF2->mb/2,pKF2->mvDepth[idx2]));

            // 得到双目观测的视差角
            cosParallaxStereo = min(cosParallaxStereo1,cosParallaxStereo2);

            // 三角化恢复3D点
            cv::Mat x3D;
            // cosParallaxRays>0 && (bStereo1 || bStereo2 || cosParallaxRays<0.9998)表明视差角正常
            // cosParallaxRays<cosParallaxStereo表明视差角很小
            // 视差角度小时用三角法恢复3D点，视差角大时用双目恢复3D点
            if(cosParallaxRays<cosParallaxStereo && cosParallaxRays>0 && (bStereo1 || bStereo2 || cosParallaxRays<0.9998))
            {
                // Linear Triangulation Method
                cv::Mat A(4,4,CV_32F);
                A.row(0) = xn1.at<float>(0)*Tcw1.row(2)-Tcw1.row(0);
                A.row(1) = xn1.at<float>(1)*Tcw1.row(2)-Tcw1.row(1);
                A.row(2) = xn2.at<float>(0)*Tcw2.row(2)-Tcw2.row(0);
                A.row(3) = xn2.at<float>(1)*Tcw2.row(2)-Tcw2.row(1);

                cv::Mat w,u,vt;
                cv::SVD::compute(A,w,u,vt,cv::SVD::MODIFY_A| cv::SVD::FULL_UV);

                x3D = vt.row(3).t();

                if(x3D.at<float>(3)==0)
                    continue;

                // Euclidean coordinates
                x3D = x3D.rowRange(0,3)/x3D.at<float>(3);
            }
            else if(bStereo1 && cosParallaxStereo1<cosParallaxStereo2)
            {
                x3D = mpCurrentKeyFrame->UnprojectStereo(idx1);                
            }
            else if(bStereo2 && cosParallaxStereo2<cosParallaxStereo1)
            {
                x3D = pKF2->UnprojectStereo(idx2);
            }
            else
                continue; //No stereo and very low parallax

            cv::Mat x3Dt = x3D.t();

            //Check triangulation in front of cameras
            // 检测深度正负
            float z1 = Rcw1.row(2).dot(x3Dt)+tcw1.at<float>(2);
            if(z1<=0)
                continue;

            float z2 = Rcw2.row(2).dot(x3Dt)+tcw2.at<float>(2);
            if(z2<=0)
                continue;

            //Check reprojection error in first keyframe
            // 计算3D点在当前关键帧下的重投影误差
            const float &sigmaSquare1 = mpCurrentKeyFrame->mvLevelSigma2[kp1.octave];
            const float x1 = Rcw1.row(0).dot(x3Dt)+tcw1.at<float>(0);
            const float y1 = Rcw1.row(1).dot(x3Dt)+tcw1.at<float>(1);
            const float invz1 = 1.0/z1;

            if(!bStereo1)
            {
                float u1 = fx1*x1*invz1+cx1;
                float v1 = fy1*y1*invz1+cy1;
                float errX1 = u1 - kp1.pt.x;
                float errY1 = v1 - kp1.pt.y;
                if((errX1*errX1+errY1*errY1)>5.991*sigmaSquare1)
                    continue;
            }
            else
            {
                float u1 = fx1*x1*invz1+cx1;
                float u1_r = u1 - mpCurrentKeyFrame->mbf*invz1;
                float v1 = fy1*y1*invz1+cy1;
                float errX1 = u1 - kp1.pt.x;
                float errY1 = v1 - kp1.pt.y;
                float errX1_r = u1_r - kp1_ur;
                if((errX1*errX1+errY1*errY1+errX1_r*errX1_r)>7.8*sigmaSquare1)
                    continue;
            }

            //Check reprojection error in second keyframe
            // 计算3D点在另一个关键帧下的重投影误差
            const float sigmaSquare2 = pKF2->mvLevelSigma2[kp2.octave];
            const float x2 = Rcw2.row(0).dot(x3Dt)+tcw2.at<float>(0);
            const float y2 = Rcw2.row(1).dot(x3Dt)+tcw2.at<float>(1);
            const float invz2 = 1.0/z2;
            if(!bStereo2)
            {
                float u2 = fx2*x2*invz2+cx2;
                float v2 = fy2*y2*invz2+cy2;
                float errX2 = u2 - kp2.pt.x;
                float errY2 = v2 - kp2.pt.y;
                if((errX2*errX2+errY2*errY2)>5.991*sigmaSquare2)
                    continue;
            }
            else
            {
                float u2 = fx2*x2*invz2+cx2;
                float u2_r = u2 - mpCurrentKeyFrame->mbf*invz2;
                float v2 = fy2*y2*invz2+cy2;
                float errX2 = u2 - kp2.pt.x;
                float errY2 = v2 - kp2.pt.y;
                float errX2_r = u2_r - kp2_ur;
                if((errX2*errX2+errY2*errY2+errX2_r*errX2_r)>7.8*sigmaSquare2)
                    continue;
            }

            //Check scale consistency
            // 检查尺度连续性
            cv::Mat normal1 = x3D-Ow1;
            float dist1 = cv::norm(normal1);

            cv::Mat normal2 = x3D-Ow2;
            float dist2 = cv::norm(normal2);

            if(dist1==0 || dist2==0)
                continue;

            const float ratioDist = dist2/dist1;
            const float ratioOctave = mpCurrentKeyFrame->mvScaleFactors[kp1.octave]/pKF2->mvScaleFactors[kp2.octave];
            if(ratioDist*ratioFactor<ratioOctave || ratioDist>ratioOctave*ratioFactor)
                continue;

            // Triangulation is succesfull
            // 三角化生成3D点成功，构造成MapPoint
            MapPoint* pMP = new MapPoint(x3D,mpCurrentKeyFrame,mpMap);

            // 为该MapPoint添加属性
            pMP->AddObservation(mpCurrentKeyFrame,idx1);            
            pMP->AddObservation(pKF2,idx2);

            mpCurrentKeyFrame->AddMapPoint(pMP,idx1);
            pKF2->AddMapPoint(pMP,idx2);

            pMP->ComputeDistinctiveDescriptors();

            pMP->UpdateNormalAndDepth();

            mpMap->AddMapPoint(pMP);

            // 将新产生的点放入检测队列
            mlpRecentAddedMapPoints.push_back(pMP);

            nnew++;
        }
    }
}

// 检查并融合当前关键帧与相邻帧（两级相邻）重复的MapPoints
void LocalMapping::SearchInNeighbors()
{
    // Retrieve neighbor keyframes
    // 获得当前关键帧在covisibility图中权重排名前nn的邻接关键帧
    // 找到当前帧一级相邻与二级相邻关键帧
    int nn = 10;
    if(mbMonocular)
        nn=20;
    const vector<KeyFrame*> vpNeighKFs = mpCurrentKeyFrame->GetBestCovisibilityKeyFrames(nn);

    vector<KeyFrame*> vpTargetKFs;
    for(vector<KeyFrame*>::const_iterator vit=vpNeighKFs.begin(), vend=vpNeighKFs.end(); vit!=vend; vit++)
    {
        KeyFrame* pKFi = *vit;
        if(pKFi->isBad() || pKFi->mnFuseTargetForKF == mpCurrentKeyFrame->mnId)
            continue;
        vpTargetKFs.push_back(pKFi);	// 一级相邻帧
        pKFi->mnFuseTargetForKF = mpCurrentKeyFrame->mnId;

        // Extend to some second neighbors
		// 二级相邻帧，5帧
        const vector<KeyFrame*> vpSecondNeighKFs = pKFi->GetBestCovisibilityKeyFrames(5);
        for(vector<KeyFrame*>::const_iterator vit2=vpSecondNeighKFs.begin(), vend2=vpSecondNeighKFs.end(); vit2!=vend2; vit2++)
        {
            KeyFrame* pKFi2 = *vit2;
            if(pKFi2->isBad() || pKFi2->mnFuseTargetForKF==mpCurrentKeyFrame->mnId || pKFi2->mnId==mpCurrentKeyFrame->mnId)
                continue;
            vpTargetKFs.push_back(pKFi2);// 存入二级相邻帧
        }
    }

    // Search matches by projection from current KF in target KFs
    ORBmatcher matcher;

    // 将当前帧的MapPoints分别与一级二级相邻帧进行融合
    vector<MapPoint*> vpMapPointMatches = mpCurrentKeyFrame->GetMapPointMatches();
    for(vector<KeyFrame*>::iterator vit=vpTargetKFs.begin(), vend=vpTargetKFs.end(); vit!=vend; vit++)
    {
        KeyFrame* pKFi = *vit;

        // 投影当前帧的地图点到pKFi中，并判断是否有重复的地图点
        // 如果地图点能匹配关键帧的特征点，并且该点有对应的地图点，那么将两个地图点合并
        // 如果地图点能匹配关键帧的特征点，并且该点没有对应的地图点，那么为该点添加地图点
        matcher.Fuse(pKFi,vpMapPointMatches);
    }

    // Search matches by projection from target KFs in current KF
    // 用于存储一级邻接和二级邻接关键帧所有MapPoints的集合
    vector<MapPoint*> vpFuseCandidates;
    vpFuseCandidates.reserve(vpTargetKFs.size()*vpMapPointMatches.size());

    // 将一级二级相邻帧的MapPoints分别与当前帧（的MapPoints）进行融合
    // 遍历每一个一级邻接和二级邻接关键帧
    for(vector<KeyFrame*>::iterator vitKF=vpTargetKFs.begin(), vendKF=vpTargetKFs.end(); vitKF!=vendKF; vitKF++)
    {
        KeyFrame* pKFi = *vitKF;

        vector<MapPoint*> vpMapPointsKFi = pKFi->GetMapPointMatches();

        // 遍历当前一级邻接和二级邻接关键帧中所有的MapPoints
        for(vector<MapPoint*>::iterator vitMP=vpMapPointsKFi.begin(), vendMP=vpMapPointsKFi.end(); vitMP!=vendMP; vitMP++)
        {
            MapPoint* pMP = *vitMP;
            if(!pMP)
                continue;
            
            // 判断MapPoints是否为坏点，或者是否已经加进集合vpFuseCandidates
            if(pMP->isBad() || pMP->mnFuseCandidateForKF == mpCurrentKeyFrame->mnId)
                continue;

            // 加入集合，并标记已经加入
            pMP->mnFuseCandidateForKF = mpCurrentKeyFrame->mnId;
            vpFuseCandidates.push_back(pMP);
        }
    }

    matcher.Fuse(mpCurrentKeyFrame,vpFuseCandidates);

    // Update points
    // 更新当前帧MapPoints的描述子，深度，观测主方向等属性
    vpMapPointMatches = mpCurrentKeyFrame->GetMapPointMatches();
    for(size_t i=0, iend=vpMapPointMatches.size(); i<iend; i++)
    {
        MapPoint* pMP=vpMapPointMatches[i];
        if(pMP)
        {
            if(!pMP->isBad())
            {
                // 在所有找到pMP的关键帧中，获得最佳的描述子
                pMP->ComputeDistinctiveDescriptors();

                pMP->UpdateNormalAndDepth();
            }
        }
    }

    // Update connections in covisibility graph

    // 步更新当前帧的MapPoints后更新与其它帧的连接关系
    // 更新covisibility图
    mpCurrentKeyFrame->UpdateConnections();
}

cv::Mat LocalMapping::ComputeF12(KeyFrame *&pKF1, KeyFrame *&pKF2)
{
    cv::Mat R1w = pKF1->GetRotation();
    cv::Mat t1w = pKF1->GetTranslation();
    cv::Mat R2w = pKF2->GetRotation();
    cv::Mat t2w = pKF2->GetTranslation();

    cv::Mat R12 = R1w*R2w.t();
    cv::Mat t12 = -R1w*R2w.t()*t2w+t1w;

    cv::Mat t12x = SkewSymmetricMatrix(t12);

    const cv::Mat &K1 = pKF1->mK;
    const cv::Mat &K2 = pKF2->mK;

    return K1.t().inv()*t12x*R12*K2.inv();
}

void LocalMapping::RequestStop()
{
    unique_lock<mutex> lock(mMutexStop);
    mbStopRequested = true;
    unique_lock<mutex> lock2(mMutexNewKFs);
    mbAbortBA = true;
}

bool LocalMapping::Stop()
{
    unique_lock<mutex> lock(mMutexStop);
    if(mbStopRequested && !mbNotStop)
    {
        mbStopped = true;
        cout << "Local Mapping STOP" << endl;
        return true;
    }

    return false;
}

bool LocalMapping::isStopped()
{
    unique_lock<mutex> lock(mMutexStop);
    return mbStopped;
}

bool LocalMapping::stopRequested()
{
    unique_lock<mutex> lock(mMutexStop);
    return mbStopRequested;
}

void LocalMapping::Release()
{
    unique_lock<mutex> lock(mMutexStop);
    unique_lock<mutex> lock2(mMutexFinish);
    if(mbFinished)
        return;
    mbStopped = false;
    mbStopRequested = false;
    for(list<KeyFrame*>::iterator lit = mlNewKeyFrames.begin(), lend=mlNewKeyFrames.end(); lit!=lend; lit++)
        delete *lit;
    mlNewKeyFrames.clear();

    cout << "Local Mapping RELEASE" << endl;
}

bool LocalMapping::AcceptKeyFrames()
{
    unique_lock<mutex> lock(mMutexAccept);
    return mbAcceptKeyFrames;
}

void LocalMapping::SetAcceptKeyFrames(bool flag)
{
    unique_lock<mutex> lock(mMutexAccept);
    mbAcceptKeyFrames=flag;
}

bool LocalMapping::SetNotStop(bool flag)
{
    unique_lock<mutex> lock(mMutexStop);

    if(flag && mbStopped)
        return false;

    mbNotStop = flag;

    return true;
}

void LocalMapping::InterruptBA()
{
    mbAbortBA = true;
}

// 关键帧剔除
void LocalMapping::KeyFrameCulling()
{
    // Check redundant keyframes (only local keyframes)
    // A keyframe is considered redundant if the 90% of the MapPoints it sees, are seen
    // in at least other 3 keyframes (in the same or finer scale)
    // We only consider close stereo points

    // 提取当前帧的共视关键帧
    vector<KeyFrame*> vpLocalKeyFrames = mpCurrentKeyFrame->GetVectorCovisibleKeyFrames();

    for(vector<KeyFrame*>::iterator vit=vpLocalKeyFrames.begin(), vend=vpLocalKeyFrames.end(); vit!=vend; vit++)
    {
        KeyFrame* pKF = *vit;
        if(pKF->mnId==0)
            continue;
        // 提取每个共视关键帧的MapPoints
        const vector<MapPoint*> vpMapPoints = pKF->GetMapPointMatches();

        int nObs = 3;
        const int thObs=nObs;
        int nRedundantObservations=0;
        int nMPs=0;

        // 遍历该局部关键帧的MapPoints，判断是否90%以上的MapPoints能被其它关键帧（至少3个）观测到
        for(size_t i=0, iend=vpMapPoints.size(); i<iend; i++)
        {
            MapPoint* pMP = vpMapPoints[i];
            if(pMP)
            {
                if(!pMP->isBad())
                {
                    if(!mbMonocular)
                    {
                        if(pKF->mvDepth[i]>pKF->mThDepth || pKF->mvDepth[i]<0)
                            continue;
                    }

                    nMPs++;
                    // MapPoints至少被三个关键帧观测到
                    if(pMP->Observations()>thObs)
                    {
                        const int &scaleLevel = pKF->mvKeysUn[i].octave;
                        const map<KeyFrame*, size_t> observations = pMP->GetObservations();
                        // 判断该MapPoint是否同时被三个关键帧观测到
                        int nObs=0;
                        for(map<KeyFrame*, size_t>::const_iterator mit=observations.begin(), mend=observations.end(); mit!=mend; mit++)
                        {
                            KeyFrame* pKFi = mit->first;
                            if(pKFi==pKF)
                                continue;
                            const int &scaleLeveli = pKFi->mvKeysUn[mit->second].octave;

                            // Scale Condition 
                            // 要求MapPoint在该局部关键帧的特征尺度大于（或近似于）其它关键帧的特征尺度
                            if(scaleLeveli<=scaleLevel+1)
                            {
                                nObs++;
                                // 已经找到三个同尺度的关键帧可以观测到该MapPoint，不用继续找了
                                if(nObs>=thObs)
                                    break;
                            }
                        }
                        // 该MapPoint至少被三个关键帧观测到
                        if(nObs>=thObs)
                        {
                            nRedundantObservations++;
                        }
                    }
                }
            }
        }

        // 该局部关键帧90%以上的MapPoints能被其它至少三个关键帧观测到，则认为是冗余关键帧
        if(nRedundantObservations>0.9*nMPs)
            pKF->SetBadFlag();
    }
}

cv::Mat LocalMapping::SkewSymmetricMatrix(const cv::Mat &v)
{
    return (cv::Mat_<float>(3,3) <<             0, -v.at<float>(2), v.at<float>(1),
            v.at<float>(2),               0,-v.at<float>(0),
            -v.at<float>(1),  v.at<float>(0),              0);
}

void LocalMapping::RequestReset()
{
    {
        unique_lock<mutex> lock(mMutexReset);
        mbResetRequested = true;
    }

    while(1)
    {
        {
            unique_lock<mutex> lock2(mMutexReset);
            if(!mbResetRequested)
                break;
        }
        //usleep(3000);
        std::this_thread::sleep_for(std::chrono::milliseconds(3));

    }
}

void LocalMapping::ResetIfRequested()
{
    unique_lock<mutex> lock(mMutexReset);
    if(mbResetRequested)
    {
        mlNewKeyFrames.clear();
        mlpRecentAddedMapPoints.clear();
        mbResetRequested=false;
    }
}

void LocalMapping::RequestFinish()
{
    unique_lock<mutex> lock(mMutexFinish);
    mbFinishRequested = true;
}

bool LocalMapping::CheckFinish()
{
    unique_lock<mutex> lock(mMutexFinish);
    return mbFinishRequested;
}

void LocalMapping::SetFinish()
{
    unique_lock<mutex> lock(mMutexFinish);
    mbFinished = true;    
    unique_lock<mutex> lock2(mMutexStop);
    mbStopped = true;
}

bool LocalMapping::isFinished()
{
    unique_lock<mutex> lock(mMutexFinish);
    return mbFinished;
}

} //namespace ORB_SLAM

按照run函数中的顺序：

1、ProcessNewKeyFrame();

• 取出最前一个关键帧，计算其BoW
• 处理其和局部地图点之间的关系：在跟踪局部图时地图点只是和帧进行了匹配，但是没有关联，这里计算平均观测方向、距离范围、最佳描述子，更新共视图，将帧插入地图中

2、MapPointCulling(); 对上一步中引入的不好的地图点进行剔除

• 实际跟踪到的帧数比预计可观测的帧数比例小于%25时，剔除
• 某点建立之后已经过了不少于2个关键帧，但是观测到该点的帧却不多，剔除
• 点建立后过了三个关键帧依然没有被剔除，则是质量好的点

3、CreateNewMapPoints();  和相邻帧通过三角化恢复一些地图点

• 获取一些共视关键帧，对于这些帧：
• 计算其和当前帧的基线与其场景深度中值的比例，比例太小，即基线太小则不考虑和这一帧进行三角化
• 计算两帧的基础矩阵并三角化
• 由视差角大小判断如何恢复点，检查深度、投影误差、尺度连续性
• 通过一系列检测后，三角化成功，为地图点添加观测属性、计算最佳描述子、更新观测方向和深度，将该点放入检测队列中，即第2步进行检查的队列

如果队列中没有待处理的关键帧：SearchInNeighbors(); 检查并融合当前关键帧与相邻帧（两级相邻）重复的地图点。

如果队列中没有待处理的关键帧，且闭环检测没有请求停止局部图，则执行一个局部BA，之后剔除冗余的关键帧。

以上是run函数的流程，也是这个线程的过程。

注：

• tracking线程中送来的关键帧放在mlNewKeyFrames变量中
• 候选待检查的地图点放在mlpRecentAddedMapPoints变量中，ProcessNewKeyFrame()函数和CreateNewMapPoints()都会向这个变量中添加地图点