matlab_rosbag is a library for reading ROS bags in Matlab. It uses the C++ ROS API to read stored messages and lets get meta-data about the bag (e.g., topic info and message definitions similar to rosmsg show and rosbag info). The library also contains methods for working with TF messages. ROS does not not need to be installed on a machine to use this library.
You can download the compiled code for Mac and Linux from github:
https://github.com/bcharrow/matlab_rosbag/releases
If you want to compile things yourself see COMPILING.md. WARNING: This library won't work at all if your machine is big-endian.
Download the library and add the base directory to your Matlab path (i.e., add the directory that contains +ros and rosbag_wrapper). You should now be able to access ros.Bag, a Matlab class which can read ROS messages on topics from a bag and return them as structs. Multiple messages are returned as cell arrays. To get an idea of how the code works go to the example directory and look at bag_example.m and tf_example.m
The fields in the structs are guaranteed to be in the same order as they are in the message definition. There are also some utilities for converting messages from structs to matrices.
NOTE: Prior to version 1.2, bags do not store message definitions. As such, matlab_rosbag won't be able to do anything with these old bags. Don't worry about this if you've been using a not terribly old version of ROS (i.e., C turtle and later).
Matlab ships with an old version of the C++ standard library that you'll need to get rid of. To do this, follow these steps from the command line:
cd /usr/local/MATLAB/<version>/sys/os/<system>
mkdir backup
mv libstdc++* backup
With some versions of Ubuntu and Matlab, the libstdc++ libraries may be located elsewhere (e.g., /usr/local/MATLAB/bin). The same steps should fix the problem.
matlab_rosbag is licensed under the BSD license.