YAC - Yet Another Calibrator

Yet Another Calibrator (YAC) is a calibration tool for calibrating RGB cameras. Specifically:

Calibrate monocular camera intrinsics
Calibrate stereo camera intrinsics and extrinsics
Calibrate mocap marker to camera extrinsics

Build

cd <your catkin workspace>
git clone https://github.com/chutsu/yac
catkin build -DCMAKE_BUILD_TYPE=Release yac yac_ros
# Note: If you don't build in release mode the optimization will be very very very slow.

Calibration Target

yac uses a grid of AprilTags, a.k.a an AprilGrid, as a calibration target.

Click here to download and print the AprilGrid. During calibration make sure the calibration target is as flat as possible.

Calibrate Monocular Camera Intrinsics

First inorder to calibrate a monocular camera we need to create the calibration configuration file. Example calib_intel_d435i.yaml configuration file:

ros:
  bag: "/data/intel_d435i/imucam-1.bag"
  cam0_topic: "/stereo/camera0/image"

calib_target:
  target_type: 'aprilgrid'  # Target type
  tag_rows: 6               # Number of rows
  tag_cols: 6               # Number of cols
  tag_size: 0.088           # Size of apriltag, edge to edge [m]
  tag_spacing: 0.3          # Ratio of space between tags to tagSize
                            # Example: tag_size=2m, spacing=0.5m
                            #          --> tag_spacing=0.25[-]

cam0:
  proj_model: "pinhole"
  dist_model: "radtan4"
  lens_hfov: 69.4
  lens_vfov: 42.5
  resolution: [640.0, 480.0]
  rate: 30.0

The above tells yac where to find the rosbag containing the calibration data, the calibration target and camera details. Now, we need to create a launch file calib_mono.launch:

<launch>
  <node pkg="yac_ros" type="calib_mono_node" name="calib_mono_node"
        required="true" output="screen">
    <param name="config_file"
           value="$(find yac_ros)/config/calib_intel_d435i.yaml" />
  </node>
</launch>

where we call the calib_mono_node and supply the calibration configuration file calib_intel_d435i.yaml. To run the calibration issue the following commands in the terminal:

source <your catkin workspace>/devel/setup.bash
roslaunch yac_ros calib_mono.launch

Calibrate Stereo Camera Intrinsics and Extrinsics

First inorder to calibrate a stereo camera we need to create the calibration configuration file. Example calib_intel_d435i.yaml configuration file:

ros:
  bag: "/data/intel_d435i/imucam-1.bag"
  cam0_topic: "/stereo/camera0/image"
  cam1_topic: "/stereo/camera1/image"

calib_target:
  target_type: 'aprilgrid'  # Target type
  tag_rows: 6               # Number of rows
  tag_cols: 6               # Number of cols
  tag_size: 0.088           # Size of apriltag, edge to edge [m]
  tag_spacing: 0.3          # Ratio of space between tags to tagSize
                            # Example: tag_size=2m, spacing=0.5m
                            #          --> tag_spacing=0.25[-]

cam0:
  proj_model: "pinhole"
  dist_model: "radtan4"
  lens_hfov: 69.4
  lens_vfov: 42.5
  resolution: [640.0, 480.0]
  rate: 30.0

cam1:
  proj_model: "pinhole"
  dist_model: "radtan4"
  lens_hfov: 69.4
  lens_vfov: 42.5
  resolution: [640.0, 480.0]
  rate: 30.0

The above tells yac where to find the rosbag containing the calibration data, the calibration target and camera details. Now, we need to create a launch file calib_stereo.launch:

<launch>
  <node pkg="yac_ros" type="calib_stereo_node" name="calib_stereo_node"
        required="true" output="screen">
    <param name="config_file"
           value="$(find yac_ros)/config/calib_intel_d435i.yaml" />
  </node>
</launch>

where we call the calib_stereo_node and supply the calibration configuration file calib_intel_d435i.yaml. To run the calibration issue the following commands in the terminal:

source <your catkin workspace>/devel/setup.bash
roslaunch yac_ros calib_stereo.launch

Calibrate Mocap Marker to Camera Extrinsics

Example calib_mocap_camera.yaml configuration file:

ros:
  train_bag: "/data/aabm/200115-vicon-train.bag"
  test_bag: "/data/aabm/200115-vicon-test.bag"
  cam0_topic: "/cam0/image"
  body0_topic: "/body0/pose"
  target0_topic: "/target0/pose"

calib_target:
  target_type: 'aprilgrid'  # Target type
  tag_rows: 6               # Number of rows
  tag_cols: 6               # Number of cols
  tag_size: 0.088           # Size of apriltag, edge to edge [m]
  tag_spacing: 0.3          # Ratio of space between tags to tagSize
                            # Example: tagSize=2m, spacing=0.5m --> tagSpacing=0.25[-]

cam0:
  resolution: [640.0, 480.0]
  proj_model: "pinhole"
  dist_model: "radtan4"
  lens_hfov: 69.4
  lens_vfov: 42.5
  rate: 30.0

LICENCE

MIT

dorianhenning/yac