从零开始的ORB_SLAM3运行

系统要求

Ubuntu18.04，未安装ros_melodic，opencv版本为4.5.1
若opencv版本不需要更改，则参考另一篇文章快速安装ORB_SLAM3

非ROS安装流程

ORB_SLAM3源码地址https://github.com/UZ-SLAMLab/ORB_SLAM3
前置需求Pangolin，OpenCV，Eigen3，DBoW2 and g2o (Included in Thirdparty folder)，Python，ROS (optional)

Pangolin下载与编译

在 ORB_SLAM3 github 网站中，作者对Pangolin的安装有以下说明：

We use Pangolin for visualization and user interface. Dowload and install instructions can be found at: https://github.com/stevenlovegrove/Pangolin.

按照 Pangolin github网站的指示安装即可。

Eigen3下载与编译

在 ORB_SLAM3 github 网站中，作者对Eigen3的安装有以下说明：

Required by g2o (see below). Download and install instructions can be found at: http://eigen.tuxfamily.org. Required at least 3.1.0.

按照网站右侧提示下载，下载后解压，找到INSTALL文件并根据提示进行安装。

Let’s call this directory ‘source_dir’ (where this INSTALL file is).
Before starting, create another directory which we will call ‘build_dir’.
Do:
cd build_dir
cmake source_dir
make install

简化版本为：

tar -zxvf eien-3.3.9.tar.gz
cd eigen-3.3.9
mkdir build
cmake ..
sudo make install

DBoW2 and g2o下载与编译

在 ORB_SLAM3 github 网站中，作者对DBoW2 and g2o的安装有以下说明：

We use modified versions of the DBoW2 library to perform place recognition and g2o library to perform non-linear optimizations. Both modified libraries (which are BSD) are included in the Thirdparty folder.

在下载的ORB_SLAM3源码中包含了 DBoW2 and g2o 两个库，并且在./build.sh文件中包含了对这两个库的编译的脚本。

OPENCV下载与编译

在 ORB_SLAM3 github 网站中，作者对opencv的安装有以下说明：

We use OpenCV to manipulate images and features. Dowload and install instructions can be found at: http://opencv.org. Required at leat 3.0. Tested with OpenCV 3.2.0.

作者使用的为opencv3.2.0，但是我在使用opencv3.2.0编译时会出现一些问题，故安装了opencv4.5.1版本。以下为一些安装建议。
从 https://github.com/opencv/opencv/releases找到opencv4.5.1版本，下载源码并解压。
https://docs.opencv.org/master/d7/d9f/tutorial_linux_install.html.按照教程安装，化简后的步骤如下：

cd opencv-4.5.1
mkdir build
cd build
cmake ..
sudo make install

默认安装地址为/usr/local，若需要改变安装地址，则需要将cmake内条语句改为。

cmake -DCMAKE_INSTALL_PREFIX=/usr/local/opencv_4.5.1 ..

若需要多核编译，则需要把make内条语句改为：（在电脑配置不好的情况下可能会死机，j后面加上数字可以指定几个线程编译）

make install -j

更多配置信息可参考https://docs.opencv.org/master/db/d05/tutorial_config_reference.html。
在其他地址安装完opencv4.5.1之后需要更改系统路径，防止在程序运行时出现寻找不到链接库的错误。

libopencv_imgproc.so.4.5: cannot open shared object file: No such file or directory

问题解决方法参考
https://blog.csdn.net/innocent_sheld/article/details/80509651
https://blog.csdn.net/jay463261929/article/details/53860593
https://blog.csdn.net/BeeGreen/article/details/104397348

ORB_SLAM3的下载与编译

在安装完前置要求之后，开始ORB_SLAM3的编译。从https://github.com/UZ-SLAMLab/ORB_SLAM3下载源码。运行目录下的build.sh文件。

git clone https://github.com/UZ-SLAMLab/ORB_SLAM3.git
cd ORB_SLAM3
chmod +x ./build.sh
./build.sh

由于安装的是opencv4.5.1，需要更改根目录下的 CMakeLists.txt 以便 cmake 编译器找到opencv的库文件。

//36行至42行
find_package(OpenCV 4.0)
if(NOT OpenCV_FOUND)
  find_package(OpenCV 3.0)
  if(NOT OpenCV_FOUND)
    message(FATAL_ERROR "OpenCV > 3.0 not found.")
  endif()
endif()

//改为
find_package(OpenCV 4.0)

数据集下载

在 ORB_SLAM3 github 网站中，作者对opencv的安装有以下说明：

EuRoC dataset was recorded with two pinhole cameras and an inertial sensor. We provide an example script to launch EuRoC sequences in all the sensor configurations.
Download a sequence (ASL format) from http://projects.asl.ethz.ch/datasets/doku.php?id=kmavvisualinertialdatasets
Open the script “euroc_examples.sh” in the root of the project. Change pathDatasetEuroc variable to point to the directory where the dataset has been uncompressed.

从http://projects.asl.ethz.ch/datasets/doku.php?id=kmavvisualinertialdatasets下载ASL Dataset Format数据集，并把数据集解压到名为MH01的文件夹内。再将euroc_examples.sh中数据集的地址更改为解压后数据集的绝对地址。

cd ~
wget http://robotics.ethz.ch/~asl-datasets/ijrr_euroc_mav_dataset/machine_hall/MH_01_easy/MH_01_easy.zip
unzip MH_01_easy.zip
mv -rf ./mav0 ~/datasets/MH01

将euroc_examples.sh的第二行进行修改：

pathDatasetEuroc='/home/s/datasets' #Example, it is necesary to change it by the dataset path

运行exampls目录下的euroc_examples.sh即可看到ORB_SLAM3的输出。

ROS安装流程

ROS_melodic的安装

roswiki中有详细的安装步骤http://wiki.ros.org/melodic/Installation/Ubuntu
其中有几点需要注意：

在更换软件源的时候可以选择中国的源，详情可见http://wiki.ros.org/ROS/Installation/UbuntuMirrors
在添加环境的环节，注意当前使用的是bash还是zsh，从而选择在.bashrc还是.zshrc中添加source
在执行sudo rosdep init时，可能需要科学上网
最重要的一点，ROS的安装教程要选择英文版而不是中文版的，中文版的安装教程并没有英文版的全面，在sudo rosdep init之前要先执行sudo apt install python-rosdep才不会报错

cv_bridge的安装

详细安装方法参考
https://zhuanlan.zhihu.com/p/347455336
https://blog.csdn.net/wakingking88/article/details/104648636

由于ROS_melodic自带opencv版本为3，故需要重新编译cv_bridge才能使得ros与opencv4.5.1兼容。

对于已经安装好 ros-melodic-desktop-full 的计算机，在执行sudo apt remove ros-melodic-cv-bridge时会导致卸载掉额外的软件包，这是我们不希望看到的。
故我们不选则执行sudo apt remove ros-melodic-cv-bridge，而是在编译好cv_bridge软件包之后需要人工把/opt/ros/melodic文件夹下的各个库文件替换掉。以下为详细步骤：

首先需要下载cv_bridge的软件包

git clone https://github.com/ros-perception/vision_opencv.git

将下载好的软件包放入工作空间中，或者重新开辟一个工作空间

//开辟工作空间
cd ~
mkdir -p catkin_ws/src
cd catkin_ws
catkin_make

//将代码放入工作空间内
cd ~/catkin_make/src
git clone https://github.com/ros-perception/vision_opencv.git

修改vision_opencv/cv_bridge/CMakeLists.txt

//由于我使用的ubuntu版本只能使用boost库1.6.5版本，python_boost库版本为3.6，故注释掉以下部分
if(NOT ANDROID)
  find_package(PythonLibs)
#  if(PYTHONLIBS_VERSION_STRING VERSION_LESS "3.8")
    # Debian Buster
#    find_package(Boost REQUIRED python37)
#  else()
    # Ubuntu Focal
    find_package(Boost REQUIRED python)
#  endif()
#else()
find_package(Boost REQUIRED)
endif()

//由于使用的是opencv4.5.1，故注释掉以下部分
set(_opencv_version 4) 
find_package(OpenCV 4 QUIET) 
#if(NOT OpenCV_FOUND)
#   message(STATUS “Did not find OpenCV 4, trying OpenCV 3”)   
#   set(_opencv_version 3) 
#endif()

修改/home/s/catkin_ws/src/vision_opencv/cv_bridge/src/module.hpp(根据自己的位置)

//将最后一段（36～40）行改为
static void do_numpy_import( )
{
    import_array( );
}

编译cv_bridge并安装库文件到/usr/local中

//传统编译方式
cd ~/catkin_ws/src/vision_opencv/cv_bridge
mkdir build
cd build
cmake ..
make
sudo make install

//ros特有编译方式
cd ~/catkin_ws
catkin_make

将cv_bridge库文件转移到ros文件夹下

sudo rm -rf /opt/ros/melodic/include/cv_bridge
sudo rm -rf /opt/ros/melodic/share/cv_bridge
sudo rm -rf /opt/ros/melodic/lib/cv_bridge
sudo cp -rf /usr/local/include/cv_bridge /opt/ros/melodic/include/cv_bridge
sudo cp -rf /usr/local/share/cv_bridge /opt/ros/melodic/share/cv_bridge
sudo cp -rf /usr/local/lib/python2.7/dist-packages/cv_bridge /opt/ros/melodic/lib/cv_bridge

ORB_SLAM3中ROS软件包的编译

首先需要添加ros软件包路径

//打开bash环境配置文件
gedit ~/.bashrc
//在文件最下方插入下面这句话(其中PATH改为真实绝对路径)
export ROS_PACKAGE_PATH=${ROS_PACKAGE_PATH}:PATH/ORB_SLAM3/Examples/ROS

修改~/git/ORB_SLAM3/Examples/ROS/ORB_SLAM3/CMakeLists文件

//修改第33到39行代码，由于使用的是opencv4.5.1，故注释掉以下部分
find_package(OpenCV 4.0 QUIET)
#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()

修改~/git/ORB_SLAM3/Examples/ROS/ORB_SLAM3/src/ros_mono文件

//修改第62行中的rostopic话题名称为/cam0/image_raw，与数据集中topic话题名称统一
    ros::Subscriber sub = nodeHandler.subscribe("/cam0/image_raw", 1, &ImageGrabber::GrabImage,&igb);

编译ros文件夹中的源码

cd ~/git/ORB_SLAM3/Examples/ROS/ORB_SLAM3
mkdir build
cd build
cmake ..
make

下载ROS bag格式数据集并运行数据集

cd ~
wget http://robotics.ethz.ch/~asl-datasets/ijrr_euroc_mav_dataset/machine_hall/MH_01_easy/MH_01_easy.bag
//终端1
roscore
//终端2
rosbag play ~/MH_01_easy.bag
//终端3
rosrun ORB_SLAM3 Mono /home/s/git/ORB_SLAM3/Vocabulary/ORBvoc.txt /home/s/git/ORB_SLAM3/Examples/ROS/ORB_SLAM3/Asus.yaml

运行成功示例

错误提示以及修正方案

cv_bridge编译过程中出现错误

CMake Error at /usr/share/cmake-3.10/Modules/FindBoost.cmake:1947 (message):
  Unable to find the requested Boost libraries.

  Boost version: 1.65.1

  Boost include path: /usr/include

  Could not find the following Boost libraries:

          boost_python37

  No Boost libraries were found.  You may need to set BOOST_LIBRARYDIR to the
  directory containing Boost libraries or BOOST_ROOT to the location of
  Boost.
Call Stack (most recent call first):
  vision_opencv/cv_bridge/CMakeLists.txt:11 (find_package)

解决方法：参考cv_bridge的安装的第三步

cv_bridge编译过程中出现错误,make 66%时出现错误

[ 66%] Building CXX object src/CMakeFiles/cv_bridge_boost.dir/module.cpp.o
In file included from /usr/include/python2.7/numpy/ndarrayobject.h:27:0,
                 from /home/nano/catkin_ws/src/cv_bridge/src/module.hpp:26,
                 from /home/nano/catkin_ws/src/cv_bridge/src/module.cpp:35:
/home/nano/catkin_ws/src/cv_bridge/src/module.hpp: In function ‘int do_numpy_import()’:
/usr/include/python2.7/numpy/__multiarray_api.h:1537:144: error: return-statement with no value, in function returning ‘int’ [-fpermissive]
 efine import_array() {if (_import_array() < 0) {PyErr_Print(); PyErr_SetString(PyExc_ImportError, "numpy.core.multiarray failed to import"); return NUMPY_IMPORT_ARRAY_RETVAL; } }
                                                                                                                                              ^
/usr/include/python2.7/numpy/__multiarray_api.h:1537:144: note: in definition of macro ‘import_array’
 efine import_array() {if (_import_array() < 0) {PyErr_Print(); PyErr_SetString(PyExc_ImportError, "numpy.core.multiarray failed to import"); return NUMPY_IMPORT_ARRAY_RETVAL; } }
                                                                                                                                              ^~~~~~
src/CMakeFiles/cv_bridge_boost.dir/build.make:62: recipe for target 'src/CMakeFiles/cv_bridge_boost.dir/module.cpp.o' failed
make[2]: *** [src/CMakeFiles/cv_bridge_boost.dir/module.cpp.o] Error 1
CMakeFiles/Makefile2:930: recipe for target 'src/CMakeFiles/cv_bridge_boost.dir/all' failed
make[1]: *** [src/CMakeFiles/cv_bridge_boost.dir/all] Error 2
Makefile:140: recipe for target 'all' failed
make: *** [all] Error 2

解决方法：参考cv_bridge的安装的第四步

ORB_SLAM3运行过程中出现错误

Failed to load image at: /Datasets/MH01/mav0/cam0/data/1403636579763555584.png
./euroc_examples.sh: 行 7: 21285 段错误               (核心已转储)

解决方法：数据集地址出现错误，修改数据集地址即可

ORB_SLAM3的ROS软件包编译过程中出现警告

usr/bin/ld: warning: libopencv_imgcodecs.so.3.2, needed by /opt/ros/melodic/lib/libcv_bridge.so, may conflict with libopencv_imgcodecs.so.4.5

解决方法：重新安装cv_bridge并重新编译ROS软件包，参考cv_bridge的安装与ROS软件包的编译

ORB_SLAM3的ROS软件包运行过程中出现错误

ORB-SLAM3 Copyright (C) 2017-2020 Carlos Campos, Richard Elvira, Juan J. Gómez, José M.M. Montiel and Juan D. Tardós, University of Zaragoza.
ORB-SLAM2 Copyright (C) 2014-2016 Raúl Mur-Artal, José M.M. Montiel and Juan D. Tardós, University of Zaragoza.
This program comes with ABSOLUTELY NO WARRANTY;
This is free software, and you are welcome to redistribute it
under certain conditions. See LICENSE.txt.

Input sensor was set to: Monocular
段错误 (核心已转储)

解决方法：运行命令出错，重新检查命令
参考https://blog.csdn.net/qq_33591712/article/details/82918981

ORB_SLAM3的ROS软件包运行过程中出现错误

ORB-SLAM3 Copyright (C) 2017-2020 Carlos Campos, Richard Elvira, Juan J. Gómez, José M.M. Montiel and Juan D. Tardós, University of Zaragoza.
ORB-SLAM2 Copyright (C) 2014-2016 Raúl Mur-Artal, José M.M. Montiel and Juan D. Tardós, University of Zaragoza.
This program comes with ABSOLUTELY NO WARRANTY;
This is free software, and you are welcome to redistribute it
under certain conditions. See LICENSE.txt.

Input sensor was set to: Monocular

Loading ORB Vocabulary. This could take a while...
Vocabulary loaded!

Creation of new map with id: 0
Creation of new map with last KF id: 0
OpenCV Error: Bad argument (Invalid pointer to file storage) in cvGetFileNodeByName, file /build/opencv-L2vuMj/opencv-3.2.0+dfsg/modules/core/src/persistence.cpp, line 863
terminate called after throwing an instance of 'cv::Exception'
  what():  /build/opencv-L2vuMj/opencv-3.2.0+dfsg/modules/core/src/persistence.cpp:863: error: (-5) Invalid pointer to file storage in function cvGetFileNodeByName

已放弃 (核心已转储)