This is a simple python package that provides two scripts to extract information, i.e. images and LiDAR scans, into offline files.
We first provide two barebone script that does the file extraction, for images and point clouds separately.
The script for image extraction will read in a user-specified ROS bag, take all the image messages on topic /camera/image_raw, and save it without alternating its encoding to a user-designated directory.
The script for cloud extraction will read in a user-specified ROS bag, take all the point cloud messages on topic /pointcloud2d or a user-specified one, and save it as .pcd format to a user-designated directory.
To use these script, simply run at the top directory level of this package folder:
python ros2_bag_to_image/extract_images_from_bag.py <path/to/bag> <output/path>
python ros2_bag_to_image/extract_pointcloud_from_bag.py <path/to/bag> <output/path>
In this Google Drive folder, we provide a collection of example ROS bags for the purpose of calibrating the camera intrinsic as well as the camera-LiDAR extrinsic, each of which contains many camera views and LiDAR scans.
Running the image extraction script mentioned above will generate an /images folder under the user-designated /output/path and extract all images in a bag into it in .png format, named using the timestamp of these images in the bag.
Running the point cloud extraction script will create the /output/path and extract all point clouds in a bag into this directory in .pcd format, following a cloud_<timestamp>.pcd naming format.
We also provide a simple ROS2 node that opens bag files from bags/, looks for the topic /camera/image_raw, then saves the image to /images under the same bag name and image index, ordered by when the images are read by the bag. While completing this operation, the node will also publish the image to /image for real-time monitoring if necessary.
To install, run the following code:
cd ~/ros2_ws/src
git clone https://github.com/ACFR-RPG/ros2_bag_to_image.git
cd ..
colcon build --packages-select ros2_bag_to_image
source install/setup.bash
To use this package, place the ROS bags in ~/ros2_ws/src/ros2_bag_to_image/bags/ and delete ~/ros2_ws/src/ros2_bag_to_image/images/. Ensure that the $HOME path variable is set correctly. Then, run the following:
ros2 run ros2_bag_to_image ros2_bag_to_image
By default, this package comes with a few ROS bags recorded from the TurtleBot3 burger robot, placed at various angles and distances relative to a checkerboard. After running, the following images will appear:
Opening one of these images will look like:
