I want to export frames from Gazebo with timestamps of the simulation time at which they are captured, like this screenshot (made in Webots)

To capture the frames, I can use the option of the rendering camera in the world file. But my frames are in greyscale and skewed, just as others reported.

I followed the instructions to use glc and it works fine. Now if I can display the simulation time in the rendering area, I'm done. I can't make up the simulation time from the image filename, because the simulation has to start paused and I don't know at what time the user unpaused it.

For rviz you can use markers to create view-oriented text. I discovered that for two years there are plans to make something similar for Gazebo, but I found nothing implemented.

I see there is a function MovableText in Ogre, the headerfile is in my case located in my ROS directory

/opt/ros/electric/stacks/simulator_gazebo/gazebo/gazebo/include/gazebo/OgreMovableText.hh

with more details of the functions online. I found an example of how it should be used in an old version of OgreCreator.cc:

383     OgreMovableText* msg = new OgreMovableText();
384     try
385     {
386       msg->Load(name, text,"Arial",0.08);
387     }
388     catch (Ogre::Exception e)
389     {
390       std::ostringstream stream;
391       stream <<  "Unable to create the text. " << e.getDescription() <<std::endl;
392       gzthrow(stream.str() );
393     }
394     msg->SetTextAlignment(OgreMovableText::H_CENTER, OgreMovableText::V_ABOVE);
395 
396     Ogre::SceneNode *textNode = OgreAdaptor::Instance()->sceneMgr->getRootSceneNode()->createChildSceneNode(std::string(name)+"_node");
397 
398     textNode->attachObject(msg);
399     textNode->translate(0, y, 0.02);

I am trying to acces this function from my realtime joint controller plugin so I include OgreAdaptor.hh, OgreCreator.hh and OgreMovableText.hh to my source and add their location to include_directories in CMakeLists.txt, and then try to reproduce the above, but I get a lot of errors related to Ogre. I don't think it is useful to post them here, because it is probably not even possible the way I'm doing this. Anybody an idea?

As far as I know, glc can only capture the rendering area of the gazebo window. If the above doesn't work, I will try to capture the whole gazebo window instead of only the rendering area with a program different from glc, then I have also the simulation time at the bottom of the pictures, which is less nice for a presentation.