如何把点云地图程序抄到你的ORBSLAM2中

---

代码从哪里抄过来？

ORBSLAM2_with_pointcloud_map

---

一、还有什么要补充的吗？

要完全补全还需要结合另外一篇文章
彩色地图和点云保存代码补全参考这里

二、操作步骤

1.引入库

CMakeLists.txt

# adding for point cloud viewer and mapper
find_package( PCL 1.7 REQUIRED )
include_directories(
${PROJECT_SOURCE_DIR}
${PROJECT_SOURCE_DIR}/include
${EIGEN3_INCLUDE_DIR}
${Pangolin_INCLUDE_DIRS}
${PCL_INCLUDE_DIRS}
)

add_definitions( ${PCL_DEFINITIONS} )
link_directories( ${PCL_LIBRARY_DIRS} )
add_library(${PROJECT_NAME} SHARED
src/System.cc
src/Tracking.cc
src/LocalMapping.cc
src/LoopClosing.cc
src/ORBextractor.cc
src/ORBmatcher.cc
src/FrameDrawer.cc
src/Converter.cc
src/MapPoint.cc
src/KeyFrame.cc
src/Map.cc
src/MapDrawer.cc
src/Optimizer.cc
src/PnPsolver.cc
src/Frame.cc
src/KeyFrameDatabase.cc
src/Sim3Solver.cc
src/Initializer.cc
src/Viewer.cc
src/pointcloudmapping.cc //增加此处
)
target_link_libraries(${PROJECT_NAME}
${OpenCV_LIBS}
${EIGEN3_LIBS}
${Pangolin_LIBRARIES}
${PROJECT_SOURCE_DIR}/Thirdparty/DBoW2/lib/libDBoW2.so
${PROJECT_SOURCE_DIR}/Thirdparty/g2o/lib/libg2o.so
${PCL_LIBRARIES}
)

2.配置文件增加一条

Examples/RGB-D/TUM1.yaml

PointCloudMapping.Resolution: 0.001//增加此处
最末尾

3.源代码增删

建立新的文件
可以直接从ORBSLAM2_with_pointcloud_map复制
include/pointcloudmapping.h

#ifndef POINTCLOUDMAPPING_H
#define POINTCLOUDMAPPING_H

#include "System.h"

#include 
#include 
#include 
#include 
#include 
using namespace ORB_SLAM2;

class PointCloudMapping
{
public:
    typedef pcl::PointXYZRGBA PointT;
    typedef pcl::PointCloud<PointT> PointCloud;

    PointCloudMapping( double resolution_ );

    // 插入一个keyframe，会更新一次地图
    void insertKeyFrame( KeyFrame* kf, cv::Mat& color, cv::Mat& depth );
    void shutdown();
    void viewer();
    void save();
protected:
    PointCloud::Ptr generatePointCloud(KeyFrame* kf, cv::Mat& color, cv::Mat& depth);

    PointCloud::Ptr globalMap;
    shared_ptr<thread>  viewerThread;

    bool    shutDownFlag    =false;
    mutex   shutDownMutex;

    condition_variable  keyFrameUpdated;
    mutex               keyFrameUpdateMutex;

    // data to generate point clouds
    vector<KeyFrame*>       keyframes;
    vector<cv::Mat>         colorImgs;
    vector<cv::Mat>         depthImgs;
    mutex                   keyframeMutex;
    uint16_t                lastKeyframeSize =0;

    double resolution = 0.04;
    pcl::VoxelGrid<PointT>  voxel;
};

#endif // POINTCLOUDMAPPING_H

include/System.h

// for point cloud viewing
#include "pointcloudmapping.h"
class PointCloudMapping;
private:
// point cloud mapping
    shared_ptr<PointCloudMapping> mpPointCloudMapping;

include/Tracking.h

// for pointcloud mapping and viewing
#include "pointcloudmapping.h"
class PointCloudMapping;
public:
    //Tracking(System* pSys, ORBVocabulary* pVoc, FrameDrawer* pFrameDrawer, MapDrawer* pMapDrawer, Map* pMap,
             //KeyFrameDatabase* pKFDB, const string &strSettingPath, const int sensor);
    Tracking(System* pSys, ORBVocabulary* pVoc, FrameDrawer* pFrameDrawer, MapDrawer* pMapDrawer, Map* pMap, shared_ptr<PointCloudMapping> pPointCloud,
             KeyFrameDatabase* pKFDB, const string &strSettingPath, const int sensor);
     cv::Mat mlmrgb;//增加此处
    cv::Mat mImDepth; 
protected:
  // for point cloud viewing
    shared_ptr<PointCloudMapping> mpPointCloudMapping;

新建如下文件
可以直接从ORBSLAM2_with_pointcloud_map复制
src/pointcloudmapping.cc

#include "pointcloudmapping.h"
#include 
#include 
#include 
#include 
#include "Converter.h"

#include 

PointCloudMapping::PointCloudMapping(double resolution_)
{
    this->resolution = resolution_;
    voxel.setLeafSize( resolution, resolution, resolution);
    globalMap = boost::make_shared< PointCloud >( );

    viewerThread = make_shared<thread>( bind(&PointCloudMapping::viewer, this ) );
}

void PointCloudMapping::shutdown()
{
    {
        unique_lock<mutex> lck(shutDownMutex);
        shutDownFlag = true;
        keyFrameUpdated.notify_one();
    }
    viewerThread->join();
}

void PointCloudMapping::insertKeyFrame(KeyFrame* kf, cv::Mat& color, cv::Mat& depth)
{
    cout<<"receive a keyframe, id = "<<kf->mnId<<endl;
    unique_lock<mutex> lck(keyframeMutex);
    keyframes.push_back( kf );
    colorImgs.push_back( color.clone() );
    depthImgs.push_back( depth.clone() );

    keyFrameUpdated.notify_one();
}
void PointCloudMapping::save()
{
	pcl::io::savePCDFile( "result.pcd", *globalMap );
	cout<<"globalMap save finished"<<endl;
}
pcl::PointCloud< PointCloudMapping::PointT >::Ptr PointCloudMapping::generatePointCloud(KeyFrame* kf, cv::Mat& color, cv::Mat& depth)
{
    PointCloud::Ptr tmp( new PointCloud() );
    // cv::imshow("123",color);
    // cv::waitKey(1);
    // point cloud is null ptr
    for ( int m=0; m<depth.rows; m+=3 )
    {
        for ( int n=0; n<depth.cols; n+=3 )
        {
            float d = depth.ptr<float>(m)[n];
            if (d < 0.01 || d>10)
                continue;
            PointT p;
            p.z = d;
            p.x = ( n - kf->cx) * p.z / kf->fx;
            p.y = ( m - kf->cy) * p.z / kf->fy;
            
            p.b = color.ptr<uchar>(m)[n*3];
            p.g = color.ptr<uchar>(m)[n*3+1];
            p.r = color.ptr<uchar>(m)[n*3+2];

            tmp->points.push_back(p);
        }
    }

    Eigen::Isometry3d T = ORB_SLAM2::Converter::toSE3Quat( kf->GetPose() );
    PointCloud::Ptr cloud(new PointCloud);
    pcl::transformPointCloud( *tmp, *cloud, T.inverse().matrix());
    cloud->is_dense = false;

    cout<<"generate point cloud for kf "<<kf->mnId<<", size="<<cloud->points.size()<<endl;
    return cloud;
}


void PointCloudMapping::viewer()
{
    pcl::visualization::CloudViewer viewer("viewer");
    while(1)
    {
        {
            unique_lock<mutex> lck_shutdown( shutDownMutex );
            if (shutDownFlag)
            {
                break;
            }
        }
        {
            unique_lock<mutex> lck_keyframeUpdated( keyFrameUpdateMutex );
            keyFrameUpdated.wait( lck_keyframeUpdated );
        }

        // keyframe is updated
        size_t N=0;
        {
            unique_lock<mutex> lck( keyframeMutex );
            N = keyframes.size();
        }

        for ( size_t i=lastKeyframeSize; i<N ; i++ )
        {
            PointCloud::Ptr p = generatePointCloud( keyframes[i], colorImgs[i], depthImgs[i] );
            *globalMap += *p;
        }
        PointCloud::Ptr tmp(new PointCloud());
        voxel.setInputCloud( globalMap );
        voxel.filter( *tmp );
        globalMap->swap( *tmp );
        viewer.showCloud( globalMap );
        cout << "show global map, size=" << globalMap->points.size() << endl;
        lastKeyframeSize = N;
    }
}

src/System.cc

System::System(const string &strVocFile, const string &strSettingsFile, const eSensor sensor,
               const bool bUseViewer):
    mSensor(sensor), mpViewer(static_cast<Viewer*>(NULL)), mbReset(false),mbActivateLocalizationMode(false),
        mbDeactivateLocalizationMode(false)
{		   //Check settings file
    cv::FileStorage fsSettings(strSettingsFile.c_str(), cv::FileStorage::READ);
    if(!fsSettings.isOpened())
    {
       cerr << "Failed to open settings file at: " << strSettingsFile << endl;
       exit(-1);
    }

    // for point cloud resolution
    float resolution = fsSettings["PointCloudMapping.Resolution"];//增加此处
---------------------------------------------------------------------
 //Create Drawers. These are used by the Viewer
    mpFrameDrawer = new FrameDrawer(mpMap);
    mpMapDrawer = new MapDrawer(mpMap, strSettingsFile);

    // Initialize pointcloud mapping
    mpPointCloudMapping = make_shared<PointCloudMapping>( resolution );//增加此处
     //Initialize the Tracking thread
    //(it will live in the main thread of execution, the one that called this constructor)
    mpTracker = new Tracking(this, mpVocabulary, mpFrameDrawer, mpMapDrawer,
                             mpMap, mpPointCloudMapping, mpKeyFrameDatabase, strSettingsFile, mSensor);//增加此处
}
void System::Shutdown()
{
    mpLocalMapper->RequestFinish();
    mpLoopCloser->RequestFinish();
    mpPointCloudMapping->shutdown();//增加此处
    }
void System::save()
{
    mpPointCloudMapping->save();
}

src/Tracking.cc

Tracking::Tracking(System *pSys, ORBVocabulary* pVoc, FrameDrawer *pFrameDrawer, MapDrawer *pMapDrawer, Map *pMap, shared_ptr<PointCloudMapping> pPointCloud, KeyFrameDatabase* pKFDB, const string &strSettingPath, const int sensor):
    mState(NO_IMAGES_YET), mSensor(sensor), mbOnlyTracking(false), mbVO(false), mpORBVocabulary(pVoc),
    mpPointCloudMapping( pPointCloud ),
    mpKeyFrameDB(pKFDB), mpInitializer(static_cast<Initializer*>(NULL)), mpSystem(pSys), mpViewer(NULL),
    mpFrameDrawer(pFrameDrawer), mpMapDrawer(pMapDrawer), mpMap(pMap), mnLastRelocFrameId(0)
void Tracking::CreateNewKeyFrame()
{
  mpLocalMapper->SetNotStop(false);

    // insert Key Frame into point cloud viewer
    mpPointCloudMapping->insertKeyFrame( pKF, this->mlmrgb, this->mImDepth );//增加此处

    mnLastKeyFrameId = mCurrentFrame.mnId;
    mpLastKeyFrame = pKF;
}

检查下方函数写法，
如果有cv::Mat imDepth = imD;
则要改写，替换成mImDepth
否则，建图结果全是散的
cv::Mat Tracking::GrabImageRGBD(
    const cv::Mat &imRGB,           //彩色图像
    const cv::Mat &imD,             //深度图像
    const double &timestamp,       //时间戳
    const vector<cv::Mat> &mask_solov2)   //solov2遮罩       
{   
    mlmrgb=imRGB; //增加此处
    mImGray = imRGB;
    mImDepth= imD;
    // cv::Mat imDepth = imD;

    // step 1：将RGB或RGBA图像转为灰度图像
    if(mImGray.channels()==3)
    {
        if(mbRGB)
            cvtColor(mImGray,mImGray,CV_RGB2GRAY);
        else
            cvtColor(mImGray,mImGray,CV_BGR2GRAY);
    }
    else if(mImGray.channels()==4)
    {
        if(mbRGB)
            cvtColor(mImGray,mImGray,CV_RGBA2GRAY);
        else
            cvtColor(mImGray,mImGray,CV_BGRA2GRAY);
    }

    // step 2 ：将深度相机的disparity转为Depth , 也就是转换成为真正尺度下的深度
    //这里的判断条件感觉有些尴尬
    //前者和后者满足一个就可以了
    //满足前者意味着,mDepthMapFactor 相对1来讲要足够大
    //满足后者意味着,如果深度图像不是浮点型? 才会执行
    //意思就是说,如果读取到的深度图像是浮点型,就不执行这个尺度的变换操作了呗? TODO 
    if((fabs(mDepthMapFactor-1.0f)>1e-5) || mImDepth.type()!=CV_32F)
        mImDepth.convertTo(  //将图像转换成为另外一种数据类型,具有可选的数据大小缩放系数
            mImDepth,            //输出图像
            CV_32F,             //输出图像的数据类型
            mDepthMapFactor);   //缩放系数

    // 步骤3：构造Frame
    mCurrentFrame = Frame(
        mImGray,                //灰度图像
        mImDepth,                //深度图像
        timestamp,              //时间戳
        mask_solov2,            //solov2遮罩
        mpORBextractorLeft,     //ORB特征提取器
        mpORBVocabulary,        //词典
        mK,                     //相机内参矩阵
        mDistCoef,              //相机的去畸变参数
        mbf,                    //相机基线*相机焦距
        mThDepth);              //内外点区分深度阈值

    // 步骤4：跟踪
    Track();

    //返回当前帧的位姿
    return mCurrentFrame.mTcw.clone();
}

---

总结