Add Disparity Map and Point Cloud Topics to the Simulation

[Antonio-Griego]

This pull request adds a launch file that starts a stereo image processing node.

• launch/stereo_image_processing.launch

This node is part of a ROS package defined in detail here: http://wiki.ros.org/stereo_image_proc

• There is a lot of details about the topics generated, frame of reference (the point cloud is based off of the left camera, for example), and other implementation details

It generates a disparity map based on the left and right camera data sensors on the robot. It also generates a point cloud using the disparity and the cameras.

Simple use instructions:

• After starting the sim, in a new terminal window, CD into the launch file directory and use the command roslaunch stereo_image_processing.launch

Specific Requests for Feedback:

• This launch file was added to the top level /launch directory. Is this appropriate? Should this be merged into a larger launch file?
• RVIZ integration: Should that be added to this pull request? Specifically, I mean should we add the disparity map output and pointcloud data to the RVIZ config file? I have left it out for now, but depending on feedback I will go back and add it in.

[Carter90]

Yes adding the changes to have the /scout_1/points2 point cloud to the rviz file would be appreciated

[Carter90]

The merge into scoots launch was good tested. Looks like we need to do some calibration here is a view of round 2's excavator's bucket

[Antonio-Griego]

I ran the updated launch script and things appear to work:

One issue I'm not familiar with is that "Task manager restarted" error. I seem to consistently get it. Is this expected behavior from something or is it a break I need to fix?

Other than that uncertainty it seems like this is good from my end.

[Antonio-Griego]

The merge into scoots launch was good tested. Looks like we need to do some calibration here is a view of round 2's excavator's bucket

Thanks for finding this,

is this calibration something that must be done before the pull request is complete? or is this a separate issue?

[Carter90]

The task restated is just that if you ran the task node before during the same simulation it will restart the state machine where it left off. So it's expected, the reason it's an "error" is if task crashes and gets restarted we want to know about it but now when we kill it.

[Carter90]

We can calibrate later, I'm happy to have data even if I mentally have to set an offset