[Documentation] [TitleIndex] [WordIndex

Only released in EOL distros:  

Goal

Bag files including camera images tend to be very large. To reduce the overall size, this packages converts the camera images in the bag file to videos and removes them from the bag file. The resulting bag file and videos can be converted back to one bag file afterwards.

Approach

sensor_msgs/CameraInfo A bag file including sensor_msgs/CameraInfo, color and depth images, both of type sensor_msgs/Image, wire_msgs/WorldEvidence and tf/tfMessage will be split into a number of video streams and a bag file. One video stream for each of the topics from some image type and a bag file for the remaining topics. In addition the package is able to merge the resulting bag file and videos back into one large bag file again.

Adapting the package to work for your own bag file with different topics can be done by simply adding or removing topics from the code.

Launching the package

Various examples can be found in the tutorials of the wire stack.

Planning

The idea is to extend this package such that it works for a bag file containing an arbitrary number of topics next to the image topics. There is, however, no timeline associated with this idea.


2024-03-16 12:56