This program will help you annotate a rosbag file by producing a result file with timestamps of the annotated events. Currently works only for Image and CompressedImage topics

It takes as input the rosbag file and a csv file if exists with annotated(or not) bound boxes from the video. The csv file holds the annotations from the video and some other useful metrics. It playbacks the rosbag file and give you some basic controls over the display and the ability to draw bound boxes on the video surface which saves them in the csv when exiting the application. A Gant is included below the video to show annotations from the video. The boxes are drown as: <HumanId::Action>

####This video player annotator is based on pyqt5 library.

labels.json file is used to add or remove any class label for annotation usage.

• OS: Ubuntu 16.04

• Python 2.7.xx

  You can check your version with: python -V command

• Ros-Kinetic: http://wiki.ros.org/kinetic/Installation/Ubuntu

• Download synaptic package manager '''sudo apt-get install''' synaptic and select the libraries below:

  1. ros-kinetic-python-qt-binding
  
  2. pyqt5-dev
  
  3. pyqt5-dev-tools
  
  4. python-pyqt5.qtmultimedia
  
  5. python-pyqt5
  

• MATPLOTLIB

  `sudo apt-get install python-matplotlib`
  

• NUMPY

  `sudo apt-get install python-numpy`
  

Run the annotator with ./rosbag_annotator_v2.py command.

Use Left Click on top-left and bottom-right to draw a new bound box. Use right click and a drop down context menu shown to annotate or delete the the specific or all bound boxes on current frame.