Contents
We utilize sensors, such as camera and lidar, in order to acquire information of our environment. Fusion of data from sensors allows you to identify objects in a scene. Cam2lidar converts data from these 2 sensors into the same coordinate system.
New in noetic
Supported hardware
Cam2lidar will work with any camera driver node satisfying the standard ROS camera interface. To be more specific, we have tested it with Intel's Realsense d435i and Econ's E-cam130 Cameras.
Tutorial
The tutorial for setting up the necessary parameters and performing the calibration can be found here.
Usage
To run the geometric_calibration node for a monocular camera set the input topics and the configuration parameters in ./config/geometric.yaml according to the tutorial and execute:
roslaunch cam2lidar geometric.launch
Accordingly, to run the temporal_calibration node for a monocular camera set the input topics and the configuration parameters in ./config/temporal.yaml according to the tutorial and execute:
roslaunch cam2lidar temporal.launch
Nodes
user_interface
user_interface is responsible for coordinating the calibration window. This node subscribes to an image raw and a camera info topic, in order to visualize the detection of the target. The calibration window shows the current images from the cameras and waits for the user to set the Distance Threshold and the number of Consequent Frames. When the user presses the Start button, the node initiated the geometric_calibration/temporal_calibration node. When sufficient number of points are collected the user clicks Calibrate and the node extracts extract the Rotation and Translation Vectors by solving the PnP problem of 2D Image points and 3D Lidar points. The output is saved in output/geometric_calibration.txt.Subscribed Topics
image_raw (sensor_msgs/Image)- raw image topic, for monocular cameras
- raw camera info topic, for monocular cameras
geometric_calibration
geometric_calibration subscribes to ROS raw image topics and Velodyne (Lidar) topic. The node detects Apriltag and the Lidar target and saves samples, in order to solve the PnP problem. The output is saved in ./output/image_points.txt and ./output/lidar_points.txt.Subscribed Topics
image_raw (sensor_msgs/Image)- raw image topic, for monocular cameras
- raw lidar topic
temporal_calibration
temporal_calibration subscribes to ROS raw image topics and Velodyne (Lidar) topic. The node detects Apriltag and the Lidar target and saves samples, in order to calculate the time difference between the lidar and the camera input. The output is calculated and saved in ./output/time_difference.txt.Subscribed Topics
image_raw (sensor_msgs/Image)- raw image topic, for monocular cameras
- raw lidar topic
