Article directory
-
- 1. Install third-party libraries
-
- 1.1 Install dependencies
- 1.2 Install Pangolin
- 1.3 Install Eigen3
- 2. Install ORB-SLAM2
- 3. Install D435i driver
-
- 3.1 Install IntelRealSense SDK2.0
- 3.2 Installation of realsense-ros
- 4. ROS runs ORB-SLAM2 in real time
-
- 4.1 Camera calibration
- 4.2 Compile ORB-SLAM2
- 4.3 Run ORB-SLAM2 in real time
ubuntu20.04
1. Install third-party libraries
1.1 Installation dependencies
sudo apt-get update sudo apt-get install git gcc g++ vim make cmake
1.2 Install Pangolin
Pangolin compressed package
(1) Install dependencies
sudo apt-get install libglew-dev libboost-dev libboost-thread-dev libboost-filesystem-dev sudo apt-get install ffmpeg libavcodec-dev libavutil-dev libavformat-dev libswscale-dev libpng-dev
(2) Configure and compile
Remember to change the downloaded file name when cd
cd Pangolin mkdir build & amp; & amp; cd build cmake -DCPP11_NO_BOOST=1 .. make sudo make install
(3) Execute the command to check whether the installation is successful
cd ../examples/HelloPangolin mkdir build & amp; & amp; cd build cmake.. make ./HelloPangolin
If the installation is successful, the following screen will pop up.
1.3 Install Eigen3
Eigen3 download address
(1) Clone source code
git clone [email protected]:eigenteam/eigen-git-mirror.git
run command
cd eigen3 mkdir build & amp; & amp; cd build cmake.. make sudo make install
(2) Copy the eigen header file to the calling folder in the root directory
sudo cp -r /usr/local/include/eigen3 /usr/include
2. Install ORB-SLAM2
ORB-SLAM2 download address
(1) Clone source code
git clone [email protected]:mirrors/raulmur/ORB_SLAM2.git
run command
cd ORB_SLAM2 chmod +x build.sh ./build.sh
chmod + ): such as make -j4/j8/jx
(2) An error occurs during compilation
error: 'usleep' was not declared in this scope
Add: before the header file corresponding to the error:
#include <unistd.h>
An error occurred:
error: static assertion failed: std::map must have the same value_type as its allocator
Put the include/LoopClosing.h file in the ORB-SLAM2 source code directory
typedef map<KeyFrame*,g2o::Sim3,std::less<KeyFrame*>,
Eigen::aligned_allocator<std::pair<const KeyFrame*, g2o::Sim3> > > KeyFrameAndPose;
change into
typedef map<KeyFrame*,g2o::Sim3,std::less<KeyFrame*>,
Eigen::aligned_allocator<std::pair<KeyFrame *const, g2o::Sim3> > > KeyFrameAndPose;
(3) Download the official TMU data set
download link
Monocular example running rgbd_dataset_freiburg1_xyz of TUM dataset
Place the downloaded rgbd_dataset_freiburg1_xyz folder in the directory of ORB-SLAM2
(4) Run monocular
./Examples/Monocular/mono_tum Vocabulary/ORBvoc.txt Examples/Monocular/TUM2.yaml rgbd_dataset_freiburg1_xyz/
3. Install D435i driver
3.1 Install IntelRealSense SDK2.0
(1) Register the publish key of the server
sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-key F6E65AC044F831AC80A06380C8B3A55A6F3EFCDE sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv-key F6E65AC044F831AC80A06380C8B3A55A6F3EFCDE
(2) Add the server to the warehouse list
sudo add-apt-repository "deb https://librealsense.intel.com/Debian/apt-repo $(lsb_release -cs) main" -u
(3) Installation library
sudo apt update sudo apt install librealsense2-dkms sudo apt install librealsense2-utils sudo apt install librealsense2-dev sudo apt install librealsense2-dbg
(4) Check whether the installation is successful
realsense-Viewer
3.2 Realsense-ros installation
(1) Create a workspace
mkdir -p ~/catkin_ws/src cd ~/catkin_ws/src/
(2) Clone source code realsense-ros
git clone https://github.com/IntelRealSense/realsense-ros.git cd realsense-ros/ git checkout `git tag | sort -V | grep -P "^2.\d + \.\d + " | tail -1` cd..
(3) Install ros package ddynamic_reconfigure
sudo apt install ros-noetic-ddynamic-reconfigure
(4) Compile
catkin_init_workspace cd.. catkin_make clean catkin_make -DCATKIN_ENABLE_TESTING=False -DCMAKE_BUILD_TYPE=Release catkin_make install
(5) Download the function package source code to the src directory
git clone https://gitcode.com/mirrors/ros-drivers/usb_cam/tree/develop/src
(6) Recompile the workspace and configure the environment
cd ~/catkin_ws/ catkin_make source ~/catkin_ws/devel/setup.bash
(7) Test camera
roslaunch usb_cam usb_cam-test.launch
Use rqt_graph to view messages
(8) If the image is not displayed, it may be due to the following reasons
a. Report an error
RLException: [usb_cam-test.launch] is neither a launch file in package [usb_cam] nor is [usb_cam] a launch file name The traceback for the exception was written to the log file
Reason: Environment not configured
source ~/catkin_ws/devel/setup.bash
If you don’t want to enter it every time, you can add this sentence to the ~/.bashrc file
sudo gedit ~/.bashrc
Save and exit, execute:
source ~/.bashrc
b. The camera displays a blurry screen
Solution: The default pixel format of the camera is mjpeg format, the pixel format defined in the launch file is yuyv format (sixth line), change it to mjpeg
However, a warning will be prompted in the terminal at this time
eprecated pixel format used, make sure you did set range correctly
Then find line 447 of the usb_cam.cpp file in the usb_cam-develop source code src file, and add before video_sws_ = sws_getContext…:
{<!-- -->
AVPixelFormat pixFormat;
switch (avcodec_context_->pix_fmt) {<!-- -->
case AV_PIX_FMT_YUVJ420P:
pixFormat = AV_PIX_FMT_YUV420P;
break;
case AV_PIX_FMT_YUVJ422P:
pixFormat = AV_PIX_FMT_YUV422P;
break;
case AV_PIX_FMT_YUVJ444P:
pixFormat = AV_PIX_FMT_YUV444P;
break;
case AV_PIX_FMT_YUVJ440P:
pixFormat = AV_PIX_FMT_YUV440P;
break;
default:
pixFormat = avcodec_context_->pix_fmt;
break;
}
avcodec_context_->pix_fmt = pixFormat;
}
Recompile the workspace and configure the environment
c. The camera screen is black / the camera recognized is not the camera we connected to
Solution: There are many virtual cameras in the computer. We need to replace the video* in the third line of usb_cam-test.launch with the camera connected by our USB.
1. Print the computer’s camera
ls /dev/video*
2. But what I printed out was a bunch of videos. In fact, there is only one built-in camera and an external USB camera on my computer.
/dev/video0 /dev/video2 /dev/video4 /dev/video6 /dev/video1 /dev/video3 /dev/video5 /dev/video7
3. So how do we know which of the bunch of cameras printed above is the real camera? Then use the v4l2-ctl command:
sudo apt install v4l-utils
4. Below we will conduct experiments one by one with the following instructions:
v4l2-ctl -d /dev/video0 --all
5. A lot of information is printed, we don’t need to worry about it, just focus on the camera ok: 
If there is a camera ok, it means that the camera is indeed real, and vice versa. As for whether it is a computer or an external connection, just change the corresponding number in the third line of video* in usb_cam-test.launch, and then test it.
4. ROS runs ORB-SLAM2 in real time
4.1 Camera calibration
1. Calibration board
2. Installation of noetic camera calibration tool
sudo apt-get install ros-noetic-camera-calibration
3. #Start camera
roslaunch usb_cam usb_cam-test.launch
4. Open a new terminal and start calibration
Pay attention to the parameter settings (modify according to the actual situation): –size 9×6 –square 0.016
rosrun camera_calibration cameracalibrator.py --size 9x6 --square 0.016 image:=/usb_cam/image_raw camera:=/usb_cam --k-coefficients=4
Move the chessboard according to x (left and right), y (up and down), size (front and back), skew (tilt), etc., until the progress bars of x, y, size, and skew all turn to green positions.
When they all turn green, press the CALIBRATE button and wait for a while to complete the calibration.
After completion, click the SAVE button to save the data to the default directory:
The terminal will display the following information:
('Wrote calibration data to', '/tmp/calibrationdata.tar.gz')
This folder tmp is displayed in the root directory (other location, computer). Double-click to enter and you will see the calibrationdata.tar.gz compressed package. When you open it, you will find pictures and file data information:
Note: The calibration result compressed package must be moved out of the default path tep folder in a timely manner, otherwise there may be no calibration results at any time if you open it again after a while.
Click the commit button to submit the data and exit the program. You can see the prompt to write the file head_camera.yaml.
(5) After finding the calibration result file, modify Asus.yaml in the Examples/ROS/ORB_SLAM2 directory according to its data.
4.2 Compile ORB-SLAM2
Copy the ORB_SLAM2 source code folder to the ~/catkin_ws/src folder, and open the terminal in the ORB_SLAM2 folder
(1) Modify the ros example source file
Add header files to all .cc files in the Examples/ROS/ORB_SLAM2/src path
#include <unistd.h>
Subscribe to the single-mono ros_mono.cc and ros_mono_ar.cc files:
ros::Subscriber sub = nodeHandler.subscribe("/camera/image_raw", 1, & amp;ImageGrabber::GrabImage, & amp;igb);
Replace with:
ros::Subscriber sub = nodeHandler.subscribe("/usb_cam/image_raw", 1, & amp;ImageGrabber::GrabImage, & amp;igb);
(2) Compile
cd ~/catkin_ws/src/ORB-SLAM2 chmod +x build.sh build_ros.sh ./build.sh
4.3 Run ORB-SLAM2 in real time
roslaunch usb_cam usb_cam-test.launch rosrun ORB_SLAM2 Mono Vocabulary/ORBvoc.txt Examples/ROS/ORB_SLAM2/Asus.yaml
