ContextQMD
Back to Mvision

lad_slam 直接法稠密点云slam 视觉里程计 安装

vSLAM/lsd_slam/install.md

latest18.1 KB
Original Source

lad_slam 直接法稠密点云slam 视觉里程计 安装

Large Scale Direct Monocular

ubuntu-install-lsd-slam

项目主页

github 代码

1. 官方编译方法 rosmake 编译

1. 依赖项：
sudo apt-get install python-rosinstall
sudo apt-get install ros-indigo-libg2o ros-indigo-cv-bridge liblapack-dev 
sudo apt-get install libblas-dev freeglut3-dev libqglviewer-dev libsuitesparse-dev libx11-dev

2. rosbuild_ws：
mkdir ~/SLAM/Code/rosbuild_ws　　　#创建 编译空间
cd ~/SLAM/Code/rosbuild_ws 
roses init . /opt/ros/indigo       #编译空间初始化

mkdir package_dir
roses set ~/SLAM/Code/rosbuild_ws/package_dir -t .      #设置
echo "source ~/SLAM/Code/rosbuild_ws/setup.bash" >> ~/.bashrc
bash

3. 下载编译:
cd package_dir
git clone https://github.com/tum-vision/lsd_slam.git lsd_slam

rosmake lsd_slam

4. 如果你需要openFabMap去闭环检测的话（可选） 需要opencv 非免费的包
在 lsd_slam_core/CMakeLists.txt  
中去掉下列四行注释即可
#add_subdirectory(${PROJECT_SOURCE_DIR}/thirdparty/openFabMap)  
#include_directories(${PROJECT_SOURCE_DIR}/thirdparty/openFabMap/include)  
#add_definitions("-DHAVE_FABMAP")  
#set(FABMAP_LIB openFABMAP )  

需要注意openFabMap需要OpenCv nonfree模块支持，OpenCV3.0以上已经不包含，作者推荐2.4.8版本
其中nonfree模块可以由以下方式安装
$ sudo add-apt-repository --yes ppa:xqms/opencv-nonfree  
$ sudo apt-get update  
$ sudo apt-get install libopencv-nonfree-dev

2. 使用catkin对LSD-SLAM进行编译

catkin_make编译 参考

后面有详细的记录

    上面的依赖项还需要安装

    下载:
mkdir -p ~/catkin_ws/src
git clone https://github.com/tum-vision/lsd_slam.git
cd lsd_slam
git checkout catkin

　　　 修改 package.xml: 对lsd_slam/lsd_slam_viewer和lsd_slam/lsd_slam_core文件夹下的package.xml中添加： <build_depend>cmake_modules</build_depend> <run_depend>cmake_modules</run_depend>

    修改 CMakeFiles.txt
对lsd_slam/lsd_slam_viewer和lsd_slam/lsd_slam_core文件夹下的CMakeFiles.txt中添加：
find_package(cmake_modules REQUIRED)
find_package(OpenCV 3.0 QUIET) #support opencv3
if(NOT OpenCV_FOUND)
   find_package(OpenCV 2.4.3 QUIET)
   if(NOT OpenCV_FOUND)
      message(FATAL_ERROR "OpenCV > 2.4.3 not found.")
   endif()
endif()

并且在所有的target_link_libraries中添加X11 ${OpenCV_LIBS}，如：
target_link_libraries(lsdslam 
${FABMAP_LIB} 
${G2O_LIBRARIES} 
${catkin_LIBRARIES} 
${OpenCV_LIBS} 
sparse cxsparse X11
)

然后开始编译：
cd ~/catkin_ws/
catkin_make

3. 数据 测试

下载测试数据474MB日志回放

解压并运行

打开一个终端:
roscoe

打开另外一个终端：
cd ~/catkin_ws/
source devel/setup.sh
rosrun lsd_slam_viewer viewer

打开另外一个终端：
cd ~/catkin_ws/
source devel/setup.sh
rosrun lsd_slam_core live_slam image:=/image_raw camera_info:=/camera_info

打开另外一个终端：
cd ~/catkin_ws/
rosbag play ~/LSD_room.bag  回放日志   即将之前的数据按话题发布

4. 使用摄像头运行LSD_SLAM

安装相机驱动

cd ~/catkin_ws/
source devel/setup.sh
cd ~/catkin_ws/src
git clone https://github.com/ktossell/camera_umd.git
cd ..
catkin_make
roscd uvc_camera/launch/
roslaunch ./camera_node.launch

添加 camera_node.launch

<launch>
  <node pkg="uvc_camera" type="uvc_camera_node" name="uvc_camera" output="screen">
    <param name="width" type="int" value="640" />
    <param name="height" type="int" value="480" />
    <param name="fps" type="int" value="30" />
    <param name="frame" type="string" value="wide_stereo" />

    <param name="auto_focus" type="bool" value="False" />
    <param name="focus_absolute" type="int" value="0" />
    <!-- other supported params: auto_exposure, exposure_absolute, brightness, power_line_frequency -->

    <param name="device" type="string" value="/dev/video1" />
    <param name="camera_info_url" type="string" value="file://$(find uvc_camera)/example.yaml" />
  </node>
</launch>

运行测试　注意官方程序默认分辨率为640*480

1. 启动ROS服务  
   打开一个窗口 运行
   roscore；

2. 启动摄像服务（USB摄像模式）
   打开另外一个窗口：
   cd ~/catkin_ws/
   source devel/setup.sh
   roslaunch uvc_camera camera_node.launch device:=/dev/video0  
   
3. 启动LSD-viewer查看点云 
   cd ~/catkin_ws/
   source devel/setup.sh
   rosrun lsd_slam_viewer viewer


4. 启动LSD-core  
   1. 数据集模式
   rosrun lsd_slam_core dataset_slam _files:=<files> _hz:=<hz> _calib:=<calibration_file>  
   _files填数据集png包的路径，_hz表示帧率，填0代表默认值，_calib填标定文件地址 
   如：
   rosrun lsd_slam_core data
   set_slam _files:=<your path>/LSD_room/images _hz:=0 _calib:=<your path>/LSD_room/cameraCalibration.cfg  

   2. USB摄像模式
   rosrun lsd_slam_core live_slam /image:=<yourstreamtopic> _calib:=<calibration_file>  
   _calib同上，/image选择视频流 
   如：
   rosrun lsd_slam_core live_slam /image:=image_raw _calib:=<your path>/LSD_room/cameraCalibration.cfg

5. 编译错误记录

1. opencv

KeyFrameDisplay.h

//#include <opencv2/core/types_c.h>
#include <opencv2/core.hpp>
#include <opencv2/core/utility.hpp>

2. sophus 错误

opt/ros/indigo/include/sophus/sim3.hpp:339:5: error: passing ‘const RxSO3Type {aka const Sophus::RxSO3Group}’ as ‘this’ argument of ‘void Sophus::RxSO3GroupBase::setScale(const Scalar&) [with Derived = Sophus::RxSO3Group; Sophus::RxSO3GroupBase::Scalar = double]’ discards qualifiers [-fpermissive]
rxso3().setScale(scale);
^
make[3]: *** [CMakeFiles/lsdslam.dir/src/DepthEstimation/DepthMap.cpp.o] Error 1
make[3]: *** [CMakeFiles/lsdslam.dir/src/SlamSystem.cpp.o] Error 1

6. catkin_make 编译 详细记录

1. 目录下有两个包 core 和 view 所以 在 lsd_slam目录下新建一个文件夹 lad_slam

把 CMakeLists.txt 和 package.xml 放入

包信息 package.xml

       =========================
        <?xml version="1.0"?>
        <package>
          <name>lsd_slam</name>
          <version>0.0.0</version>
          <description>
            Large-Scale Direct Monocular SLAM
         </description>

          <author>Jakob Engel</author>
          <maintainer email="[email protected]">Jakob Engel</maintainer>
          <license>see http://vision.in.tum.de/lsdslam </license>
          <url>http://vision.in.tum.de/lsdslam</url>

          <license>TODO</license>
          <buildtool_depend>catkin</buildtool_depend>
          <run_depend>lsd_slam_core</run_depend>
          <run_depend>lsd_slam_viewer</run_depend>

          <export>
            <metapackage/>
          </export>
        </package>
        ====================

CMakeLists.txt

        ====================
        cmake_minimum_required(VERSION 2.8.3)
        project(lsd_slam)
        find_package(catkin REQUIRED)
        catkin_metapackage()
        ====================

2. lsd_slam_viewer

包信息 package.xml

    ==============================
    <package>
    <name>lsd_slam_viewer</name>
    <version>0.0.0</version>  

    <description>
     3D Viewer for LSD-SLAM.
    </description>

    <author>Jakob Engel</author>
    <maintainer email="[email protected]">Jakob Engel</maintainer>
    <license>see http://vision.in.tum.de/lsdslam</license>
    <url>http://vision.in.tum.de/lsdslam</url>

    <buildtool_depend>catkin</buildtool_depend>
    <build_depend>cv_bridge</build_depend>
    <build_depend>dynamic_reconfigure</build_depend>
    <build_depend>sensor_msgs</build_depend>
    <build_depend>roscpp</build_depend>
    <build_depend>roslib</build_depend>
    <build_depend>rosbag</build_depend>
    <build_depend>message_generation</build_depend>
    <build_depend>cmake_modules</build_depend>

    <run_depend>cmake_modules</run_depend>
    <run_depend>cv_bridge</run_depend>
    <run_depend>dynamic_reconfigure</run_depend>
    <run_depend>sensor_msgs</run_depend>
    <run_depend>roscpp</run_depend>
    <run_depend>roslib</run_depend>
    <run_depend>rosbag</run_depend>

    </package>

   ======================================

CMakeLists.txt

   ======================================
    cmake_minimum_required(VERSION 2.4.6)
    project(lsd_slam_viewer)

    # Set the build type. Options are:
    #  Coverage : w/ debug symbols, w/o optimization, w/ code-coverage
    #  Debug : w/ debug symbols, w/o optimization
    #  Release : w/o debug symbols, w/ optimization
    #  RelWithDebInfo : w/ debug symbols, w/ optimization
    #  MinSizeRel : w/o debug symbols, w/ optimization, stripped binaries
    set(CMAKE_BUILD_TYPE Release)

    ADD_SUBDIRECTORY(${PROJECT_SOURCE_DIR}/thirdparty/Sophus)

    set(CMAKE_MODULE_PATH ${CMAKE_MODULE_PATH} ${PROJECT_SOURCE_DIR}/cmake)

    find_package(catkin REQUIRED COMPONENTS
    cv_bridge
    dynamic_reconfigure
    sensor_msgs
    roscpp
    rosbag
    message_generation
    roslib
    )
    #find_package(OpenCV 2.4.3 REQUIRED)
    find_package(OpenCV 3.0 QUIET) #support opencv3
    if(NOT OpenCV_FOUND)
    find_package(OpenCV 2.4.3 QUIET)
    if(NOT OpenCV_FOUND)
    message(FATAL_ERROR "OpenCV > 2.4.3 not found.")
    endif()
    endif()
    find_package(cmake_modules REQUIRED)

    find_package(OpenGL REQUIRED)
    set(QT_USE_QTOPENGL TRUE)
    set(QT_USE_QTXML TRUE)
    find_package(QGLViewer REQUIRED)
    find_package(Eigen REQUIRED)
    find_package(OpenCV REQUIRED)
    find_package(Boost REQUIRED COMPONENTS thread)

    include_directories(${QGLVIEWER_INCLUDE_DIR}
        ${catkin_INCLUDE_DIRS} 
        ${EIGEN_INCLUDE_DIR}
        ${QT_INCLUDES} )

    # SSE flags
    set(CMAKE_CXX_FLAGS
    "${CMAKE_CXX_FLAGS} -march=native -Wall -std=c++0x"
    )

    add_message_files(DIRECTORY msg FILES keyframeMsg.msg keyframeGraphMsg.msg)
    generate_messages(DEPENDENCIES)

    generate_dynamic_reconfigure_options(
    cfg/LSDSLAMViewerParams.cfg
    )


    # Sources files
    set(SOURCE_FILES         
    src/PointCloudViewer.cpp
    src/KeyFrameDisplay.cpp
    src/KeyFrameGraphDisplay.cpp
    src/settings.cpp
    src/keyboard/keyboard.cc# keyboard
    src/robot/robot.cc# robor model
    src/serial/serial.cc# communication
    )

    set(HEADER_FILES     
    src/PointCloudViewer.h
    src/KeyFrameDisplay.h
    src/KeyFrameGraphDisplay.h
    src/settings.h
    src/keyboard/keyboard.h
    src/robot/robot.h
    src/serial/serial.h
    )

    include_directories(
    ${PROJECT_SOURCE_DIR}/thirdparty/Sophus
    )  

    set(LIBS
    ${QGLViewer_LIBRARIES}
    ${QGLVIEWER_LIBRARY} 
    ${catkin_LIBRARIES}
    ${Boost_LIBRARIES}
    ${QT_LIBRARIES}
    GL glut GLU X11
    )

    add_executable(viewer src/main_viewer.cpp ${SOURCE_FILES} ${HEADER_FILES})
    target_link_libraries(viewer ${LIBS}
         )
    add_executable(videoStitch src/main_stitchVideos.cpp)
    target_link_libraries(videoStitch ${OpenCV_LIBS} ${LIBS})

    #add_executable(videoStitch src/main_stitchVideos.cpp)
    #target_link_libraries(viewer ${QGLViewer_LIBRARIES}
    #			     ${QGLVIEWER_LIBRARY}
    #			     ${catkin_LIBRARIES}
    #			     ${QT_LIBRARIES}
    #			     GL glut GLU
    #		      )

  =============================

3. lsd_slam_core

包信息 package.xml

     ========================
    <?xml version="1.0"?>
    <package>
      <name>lsd_slam_core</name>
      <version>0.0.0</version>
      <description>
         Large-Scale Direct Monocular SLAM
      </description>

      <author>Jakob Engel</author>
      <maintainer email="[email protected]">Jakob Engel</maintainer>
      <license>see http://vision.in.tum.de/lsdslam </license>
      <url>http://vision.in.tum.de/lsdslam</url>

      <buildtool_depend>catkin</buildtool_depend>
      <build_depend>cv_bridge</build_depend>
      <build_depend>dynamic_reconfigure</build_depend>
      <build_depend>sensor_msgs</build_depend>
      <build_depend>roscpp</build_depend>
      <build_depend>lsd_slam_viewer</build_depend>
      <build_depend>rosbag</build_depend>
      <build_depend>eigen</build_depend>
      <build_depend>suitesparse</build_depend>
      <build_depend>libg2o</build_depend>
      <build_depend>cmake_modules</build_depend>

      <run_depend>cmake_modules</run_depend>
      <run_depend>cv_bridge</run_depend>
      <run_depend>dynamic_reconfigure</run_depend>
      <run_depend>sensor_msgs</run_depend>
      <run_depend>roscpp</run_depend>
      <run_depend>lsd_slam_viewer</run_depend>
      <run_depend>rosbag</run_depend>
      <run_depend>eigen</run_depend>
      <run_depend>suitesparse</run_depend>
      <run_depend>libg2o</run_depend>

    </package> 
    ===============

CMakeLists.txt

    ===============
    cmake_minimum_required(VERSION 2.8.7)
    project(lsd_slam_core)

    set(CMAKE_BUILD_TYPE Release)

    find_package(catkin REQUIRED COMPONENTS
      cv_bridge
      dynamic_reconfigure
      sensor_msgs
      image_transport
      roscpp
      rosbag
    )

    #find_package(OpenCV 2.4.3 REQUIRED)
    find_package(OpenCV 3.0 QUIET) #support opencv3
    if(NOT OpenCV_FOUND)
       find_package(OpenCV 2.4.3 QUIET)
       if(NOT OpenCV_FOUND)
          message(FATAL_ERROR "OpenCV > 2.4.3 not found.")
       endif()
    endif()

    find_package(cmake_modules REQUIRED)

    find_package(Eigen3 REQUIRED)
    find_package(X11 REQUIRED)
    include(cmake/FindG2O.cmake)
    include(cmake/FindSuiteParse.cmake)

    message("-- CHOLMOD_INCLUDE_DIR : " ${CHOLMOD_INCLUDE_DIR})
    message("-- CSPARSE_INCLUDE_DIR : " ${CSPARSE_INCLUDE_DIR})
    message("-- G2O_INCLUDE_DIR : " ${G2O_INCLUDE_DIR})

    # FabMap
    # uncomment this part to enable fabmap
    #add_subdirectory(${PROJECT_SOURCE_DIR}/thirdparty/openFabMap)
    #include_directories(${PROJECT_SOURCE_DIR}/thirdparty/openFabMap/include)
    #add_definitions("-DHAVE_FABMAP")
    #set(FABMAP_LIB openFABMAP )

    generate_dynamic_reconfigure_options(
      cfg/LSDDebugParams.cfg
      cfg/LSDParams.cfg
    )

    catkin_package(
      LIBRARIES lsdslam
      DEPENDS Eigen SuiteSparse
      CATKIN_DEPENDS libg2o 
    )

    # SSE flags
    add_definitions("-DUSE_ROS")
    add_definitions("-DENABLE_SSE")

    # Also add some useful compiler flag
    set(CMAKE_CXX_FLAGS
       "${CMAKE_CXX_FLAGS} -march=native -Wall -std=c++0x"
    ) 
    # Set source files
    set(lsd_SOURCE_FILES
      ${PROJECT_SOURCE_DIR}/src/DataStructures/Frame.cpp
      ${PROJECT_SOURCE_DIR}/src/DataStructures/FramePoseStruct.cpp
      ${PROJECT_SOURCE_DIR}/src/DataStructures/FrameMemory.cpp
      ${PROJECT_SOURCE_DIR}/src/SlamSystem.cpp
      ${PROJECT_SOURCE_DIR}/src/LiveSLAMWrapper.cpp
      ${PROJECT_SOURCE_DIR}/src/DepthEstimation/DepthMap.cpp
      ${PROJECT_SOURCE_DIR}/src/DepthEstimation/DepthMapPixelHypothesis.cpp
      ${PROJECT_SOURCE_DIR}/src/util/globalFuncs.cpp
      ${PROJECT_SOURCE_DIR}/src/util/SophusUtil.cpp
      ${PROJECT_SOURCE_DIR}/src/util/settings.cpp
      ${PROJECT_SOURCE_DIR}/src/util/Undistorter.cpp
      ${PROJECT_SOURCE_DIR}/src/Tracking/Sim3Tracker.cpp
      ${PROJECT_SOURCE_DIR}/src/Tracking/Relocalizer.cpp
      ${PROJECT_SOURCE_DIR}/src/Tracking/SE3Tracker.cpp
      ${PROJECT_SOURCE_DIR}/src/Tracking/least_squares.cpp
      ${PROJECT_SOURCE_DIR}/src/Tracking/TrackingReference.cpp
      ${PROJECT_SOURCE_DIR}/src/IOWrapper/Timestamp.cpp
      ${PROJECT_SOURCE_DIR}/src/GlobalMapping/FabMap.cpp
      ${PROJECT_SOURCE_DIR}/src/GlobalMapping/KeyFrameGraph.cpp
      ${PROJECT_SOURCE_DIR}/src/GlobalMapping/g2oTypeSim3Sophus.cpp
      ${PROJECT_SOURCE_DIR}/src/GlobalMapping/TrackableKeyFrameSearch.cpp
    )
    set(SOURCE_FILES
      ${lsd_SOURCE_FILES}
      ${PROJECT_SOURCE_DIR}/src/IOWrapper/ROS/ROSImageStreamThread.cpp
      ${PROJECT_SOURCE_DIR}/src/IOWrapper/ROS/ROSOutput3DWrapper.cpp
      ${PROJECT_SOURCE_DIR}/src/IOWrapper/OpenCV/ImageDisplay_OpenCV.cpp
    )
    include_directories(
      include
      ${EIGEN3_INCLUDE_DIR}
      ${PROJECT_SOURCE_DIR}/src
      ${PROJECT_SOURCE_DIR}/thirdparty/Sophus
      ${CSPARSE_INCLUDE_DIR} #Has been set by SuiteParse
      ${CHOLMOD_INCLUDE_DIR} #Has been set by SuiteParse
    )
    set(LIBS
    ${catkin_LIBRARIES}
    ${G2O_LIBRARIES} 
    ${OpenCV_LIBS}
    ${EIGEN3_LIBS}
    )
    # build shared library.
    add_library(lsdslam SHARED ${SOURCE_FILES})
    target_link_libraries(lsdslam ${FABMAP_LIB} ${LIBS} csparse cxsparse 
    X11
    )
    #rosbuild_link_boost(lsdslam thread)
    # build live ros node
    add_executable(live_slam src/main_live_odometry.cpp)
    target_link_libraries(live_slam lsdslam ${LIBS}
    X11
    )
    # build image node
    add_executable(dataset src/main_on_images.cpp)
    target_link_libraries(dataset lsdslam ${LIBS}
    X11
    )
    # TODO add INSTALL

    ============