ORB-SLAM3安装运行
运行环境:Ubuntu18.04
安装依赖
下载ORB-SLAM3源码
git clone https://github.com/UZ-SLAMLab/ORB_SLAM3.git ORB_SLAM3
Pangolin安装
git clone https://github.com/stevenlovegrove/Pangolin.git
根据github上说明,安装Pangolin所需要的依赖
- C++11
- OpenGL (Desktop / ES / ES2)
- (lin)
sudo apt install libgl1-mesa-dev
- (lin)
- Glew
- (deb)
sudo apt install libglew-dev
- (deb)
- CMake (for build environment)
- (deb)
sudo apt install cmake
- (deb)
Recommended Dependencies
- Python2 / Python3, for drop-down interactive console
- (deb)
sudo apt install libpython2.7-dev
- (deb)
- Wayland
- pkg-config:
sudo apt install pkg-config
- Wayland and EGL:
sudo apt install libegl1-mesa-dev libwayland-dev libxkbcommon-dev wayland-protocols
- pkg-config:
sudo apt-get install cmake libeigen3-dev libsuitesparse-dev qtdeclarative5-dev qt5-qmake libqglviewer-dev //g2o的依赖项
sudo apt-get install libboost-dev //DBoW2的依赖项
编译安装Pangolin
cd Pangolin
mkdir build
cd build
cmake ..
cmake --build .
OpenCV(要求>3.0) 安装步骤省略
安装boost库
官方网站为https://www.boost.org/ ,下载后并解压,执行 sudo ./bootstrap.sh
和 sudo ./b2 install
安装libssl-dev sudo apt-get install libssl-dev
ORB-SLAM3编译和安装
cd ORB_SLAM3
chmod +x build.sh
./build.sh
直接执行编译脚本可能比较吃内存,过程缓慢,也可以逐条执行build.sh中的指令
# build.sh中的内容如下
echo "Configuring and building Thirdparty/DBoW2 ..."
cd Thirdparty/DBoW2
mkdir build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make -j
cd ../../g2o
echo "Configuring and building Thirdparty/g2o ..."
mkdir build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make -j
cd ../../../
echo "Uncompress vocabulary ..."
cd Vocabulary
tar -xf ORBvoc.txt.tar.gz
cd ..
echo "Configuring and building ORB_SLAM3 ..."
mkdir build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
make -j
编译错误问题
-
Eigen版本问题
修改ORB_SLAM3/Thirdparty/g2o/CMakeLists.txt 中版本号
将第72行 FIND_PACKAGE(Eigen3 3.1.0 REQUIRED)
改为 FIND_PACKAGE(Eigen3 REQUIRED)
-
make[2]: *** [CMakeFiles/ORB_SLAM3.dir/build.make:89:CMakeFiles/ORB_SLAM3.dir/src/LocalMapping.cc.o] 错误 1 make[2]: *** [CMakeFiles/ORB_SLAM3.dir/build.make:349:CMakeFiles/ORB_SLAM3.dir/src/CameraModels/KannalaBrandt8.cpp.o] 错误 1
在报错的文件里,如KannalaBrandt8.cpp,LocalMapping.cc
在include之后加上如下代码
namespace cv { template<typename _Tp, int m, int n> static inline Matx<_Tp, m, n> operator / (const Matx<_Tp, m, n> &a, float alpha) { return Matx<_Tp, m, n>(a, 1.f / alpha, Matx_ScaleOp()); } }
数据集测试
使用TUM数据集,https://vision.in.tum.de/data/datasets/rgbd-dataset/download
单目
./Examples/Monocular/mono_tum_vi Vocabulary/ORBvoc.txt Examples/Monocular/TUM_512.yaml ./data/dataset-corridor5_512_16/mav0/cam0/data Examples/Monocular/TUM_TimeStamps/dataset-corridor5_512.txt dataset-corridor5_512_mono# 其中数据集路径随便指定,这里将数据集放在当前路径下data文件夹里
RGB-D
./Examples/Monocular/mono_tum Vocabulary/ORBvoc.txt Examples/Monocular/TUM1.yaml ./data/rgbd_dataset_freiburg1_desk
运行截图如下
在线测试
在线测试需要安装ROS,请自行百度安装```
首先编译
-
将源码中的 Examples/ROS/ORB_SLAM3 路径添加到ROS_PACKAGE_PATH环境变量中
# 打开 .bashrc file:gedit ~/.bashrc# 添加路径 , # 其中PATH是本机的具体路径export ROS_PACKAGE_PATH=${ROS_PACKAGE_PATH}:“PATH”/ORB_SLAM3/Examples/ROS
-
执行build_ros.sh文件
chmod +x build_ros.sh./build_ros.sh
如果编译报错,大概就是 /ros/noetic/share 这个路径没有ORB_SLAM3的软连接,执行如下命令即可解决
sudo ln -s "PATH"/ORB_SLAM3/Examples/ROS/ORB_SLAM3 /opt/ros/noetic/share/ORB_SLAM3
在线运行需要安装usb_cam软件包
sudo apt-get install ros-noetic-usb-cam
使用笔记本摄像头运行ORB-SLAM3
首先修改源码,进入目录/ORB_SLAM3/Examples/ROS/ORB_SLAM3/src,修改ros_mono.cc文件
// 其中/usb_cam/image_raw为本机摄像头的topic,
// 使用roslaunch usb_cam usb_cam-test.launch测试摄像头时,
// 可用rostopic list查看所有topic
ros::Subscriber sub = nodeHandler.subscribe("/usb_cam/image_raw", 1, &ImageGrabber::GrabImage,&igb);
分别打开三个终端依次运行如下命令
roscore
roslaunch usb_cam usb_cam-test.launch
# 以单目摄像头运行ORB-SLAM3
rosrun ORB_SLAM3 Mono /home/jxf/ORB_SLAM3/Vocabulary/ORBvoc.txt /home/jxf/ORB_SLAM3/Examples/ROS/ORB_SLAM3/Asus.yaml
使用外接鱼眼摄像头运行ORB-SLAM3
本次实验中,使用210度鱼眼摄像头(购于某宝,100多rmb)
电脑连接usb摄像头后,首先进入目录 /opt/ros/melodic/share/usb_cam/launch
修改usb_cam-test.launch 文件,将其中/dev/video1(笔记本自带摄像头)修改为/dev/video2,
然后进入目录/ORB_SLAM3/Examples/ROS/ORB_SLAM3/src
,修改ros_mono.cc文件
//其中/usb_cam/image_raw为本机摄像头的topic,
//使用roslaunch usb_cam usb_cam-test.launch测试摄像头时,
//可用rostopic list查看所有topic
ros::Subscriber sub = nodeHandler.subscribe("/usb_cam/image_raw", 1, &ImageGrabber::GrabImage,&igb);
然后就可以运行了,分别打开三个终端依次运行如下命令
roscore
roslaunch usb_cam usb_cam-test.launch
# 以单目摄像头运行ORB-SLAM3
rosrun ORB_SLAM3 Mono /home/jxf/ORB_SLAM3/Vocabulary/ORBvoc.txt /home/jxf/ORB_SLAM3/Examples/ROS/ORB_SLAM3/Asus.yaml
运行截图如下
如果ROS运行ORB-SLAM3报错Segmentation fault (core dumped)
,则可能是因为
ORB_SLAM3/Thirdparty/DBoW2/CMakeLists.txt 和ORB_SLAM3/CMakeLists.txt指定的是opencv 4. 但是ORB_SLAM3/Examples/ROS/ORB_SLAM3/CMakeLists.txt中指定的却是opencv 3,修改前者CMakeLists.txt指定opencv版本为3,
如 set(OpenCV_DIR "/usr/local/opencv/3.4.11/share/OpenCV")
,最好修改所有的CMakeList,这样会使OpenCV版本保持一致