Calibration of monocular camera and binocular camera with kalibr calibration tool

1. Download and compile

1.1 Install dependencies , If there's a mistake in the middle , You can install the following components in multiple times , Corresponding ros The version should also be modified , My own is melodic

sudo apt-get install python-setuptools python-rosinstall ipython libeigen3-dev libboost-all-dev doxygen libopencv-dev
sudo apt-get install libopencv-dev ros-kinetic-vision-opencv ros-kinetic-image-transport-plugins ros-kinetic-cmake-modules python-software-properties software-properties-common libpoco-dev python-matplotlib python-scipy python-git python-pip ipython libtbb-dev libblas-dev liblapack-dev python-catkin-tools libv4l-dev

pip install python-igraph

1.2 compile

mkdir -p ~/kalibr_ws/src
cd ~/kalibr_ws/src
git clone --recursive https://github.com/ori-drs/kalibr

cd ~/kalibr_ws
source /opt/ros/noetic/setup.bash
catkin init
catkin config --extend /opt/ros/noetic
catkin config --merge-devel # Necessary for catkin_tools >= 0.4.
catkin config --cmake-args -DCMAKE_BUILD_TYPE=Release

catkin build -DCMAKE_BUILD_TYPE=Release -j4

problem ： Because my computer has anaconda3,anaconda3 Its own python It may affect kalibr, It seems that because kalibr Need analysis rosbag package , So the python2, Leading to an error report
Failed to load Python extension for LZ4 support. LZ4 compression will not be availabl
I'm going to cry ...

terms of settlement ：
stay kalibr_workspace/src/kalibr/aslam_offline_calibration/kalibr/python/kalibr_calibrate_cameras.py In file
change "#!/usr/bin/env python" become "#!/usr/bin/env python2"

2. Data preparation

2.1 bag Data recording

Will be recorded BAG Record the calibration file again , Remove bad pictures , At the same time bag The frequency of is changed to 4hz
Open two terminals respectively , Enter the following command and run , wait for topic The input of

rosrun topic_tools throttle messages /cam_stereo_left/csi_cam/image_raw 4.0 /left
rosrun topic_tools throttle messages /cam_stereo_right/csi_cam/image_raw 4.0 /right

Open a new terminal , For recording video

rosbag record -O stereo_calibra.bag /left /right

Open a new terminal , Used to play the original data set

rosbag play 0620_1280x720.bag

2.2 Calibration board parameter file

Write one yourself ,checkboard_8x6_10x10cm.yaml

target_type: 'checkerboard'
targetCols: 8
targetRows: 6
rowSpacingMeters: 0.1
colSpacingMeters: 0.1

target_type： Is the type of calibration plate , The checkerboard used here
targetCols： Number of corner points in the direction of calibration column
targetRows： Number of corners in the direction of calibration
rowSpacingMeters： The width of each chessboard , Unit meter
colSpacingMeters： The height of each chessboard , Unit meter

The configuration files corresponding to the other two calibration boards are attached
“ QR code calibration board ”

target_type: 'aprilgrid' #gridtype
tagCols: 6               #number of apriltags
tagRows: 6               #number of apriltags
tagSize: 0.088           #size of apriltag, edge to edge [m]
tagSpacing: 0.3          #ratio of space between tags to tagSize
                         #example: tagSize=2m, spacing=0.5m --> tagSpacing=0.25[-]

“ Circular calibration plate ”

target_type: 'circlegrid'  #gridtype
targetCols: 6              #number of circles (cols)
targetRows: 7              #number of circles (rows)
spacingMeters: 0.02        #distance between circles [m]
asymmetricGrid: False      #use asymmetric grid (opencv) [bool]

Calibration plate pdf Generate command ：
Modify as needed --nx,--ny,--tsize,--tspace

rosrun kalibr kalibr_create_target_pdf --type apriltag --nx 6 --ny 6 --tsize 0.02 --tspace 0.3

3. Start calibration

Open the terminal in the workspace directory

source YOUR_PATH/devel/setup.bash
rosrun kalibr kalibr_calibrate_cameras --bag /home/lusx/data/stereo_calibra.bag --topics /left /right --models pinhole-radtan pinhole-radtan --target src/Kalibr/config/checkboard_8x6_10x10cm.yaml

--bag: Name of calibration data 
--topics: Left and right eye cameras topic
--models: Left and right camera model 
#pinhole-radtan:  The most common pinhole model + Brownian distortion model ,  For most angles less than 120 The camera ,  The distortion parameter includes radial distortion k1,k2 And tangential distortion p1,p2;  If the camera distortion is not very serious , This model is basically OK ;  Such as my DFOV by 150 The camera ,  You can also use this and the distortion removal effect is very good ;
# pinhole-equi: Pinhole Model ＋ Equidistant distortion model , That is to say ＫＢ The type of model you need to select , The model is also widely used , Most fish eye lenses are also OK , Be careful 8 Parametric ＫＢ The distortion parameters of the model are k1,k2,k3,k4, Although it is also four numbers , But unlike the previous model , There are only radial distortion parameters , Without tangential distortion tangential distortion, The corresponding formulas for projection are also different ; At the same time, it's also opencv in cv::fisheye The model used ;
--target: Calibration board parameter configuration file 

Report errors ：
Cameras are not connected through mutual observations, please check the dataset. Maybe adjust the approx. sync. tolerance.

 The reason is that the two cameras are out of sync , It can increase the acquisition frequency to 20hz, You can also increase the synchronization tolerance at the same time  --bag-from-to 5 35 --approx-sync 0.04,
 The complete parameters are as follows ：
rosrun kalibr kalibr_calibrate_cameras --bag /home/lusx/data/stereo_calibra.bag --bag-from-to 5 100 --topics /left /right --models pinhole-radtan pinhole-radtan --target src/Kalibr/config/checkboard_8x6_10x10cm.yaml --approx-sync 0.04

3.3 Calibration results

Generate 3 File , The parameters can be in stereo_calibra-camchain.yaml see , Details can be found in stereo_calibra-report-cam.pdf View in

stereo_calibra-camchain.yaml Content ：

4 Parameter optimization

However, the internal and external parameters obtained from the calibration have a stretch , I don't know if there is a problem with the original data , Then the baseline of the two cameras was not found bf, Continue to optimize ..