Fix multiple reflectance maps and improve performance

[iche033]

Changes

• Fixes issue when there are multiple objects in the scene with different reflectance maps by creating a unique reflectance material for each reflectance map

• Improves performance by disabling shadows during reflectance render updates.

Testing

Build this branch with gazebo_ros_pkgs noetic-devel branch

Here is a reflectance_test.world for testing. The world consists of 3 depth cameras with depth, normals, reflectance output, and 100 boxes in front of the depth cameras.

For verifying multiple reflectance maps, launch gazebo with the reflectance_test.world:

roslaunch gazebo_ros empty_world.launch verbose:=true world_name:=<path_to_reflectance_test.world>

launch rqt_image_viewer to see data from the reflectance topic:

rqt_image_view /depth_cam/reflectance

You should see two different reflectance maps rendered side by side. Before the changes in the PR, I see two identical reflectances maps.

For testing performance, launch the reflectance_test.world, and see the frame rates using rostopic hz. Subscribe to all 3 depth cameras:

# you can do this in 3 terminals
rostopic hz /depth_cam/image_raw
rostopic hz /depth_cam2/image_raw
rostopic hz /depth_cam3/image_raw

The target framerate is set to 30Hz but the cameras do not hit the target frame rate (which is what we want for testing so we can see the differences in performance). These are the frame rates I get on my laptop before and after the change:

Before:

• depth;normals;reflectance: 10Hz
• depth;normals: 19Hz
• depth;reflectance: 14Hz

After:

• depth;normals;reflectance: 15Hz
• depth;normals: 19Hz
• depth;reflectance: 21Hz

In two cases that refletance is enabled, the frame rate improves by 5~7Hz

Signed-off-by: Ian Chen ichen@osrfoundation.org

[iche033]

cc @ahcorde

[ahcorde]

On my desktop
Before:

• depth;normals;reflectance: 15Hz

After:

• depth;normals;reflectance: 23Hz

[ahcorde]

Suggested change

	* [Pull request ]()
	* [Pull request](https://github.com/osrf/gazebo/pull/2742/)

[iche033]

updated. e6a2f94

[ahcorde]

LGTM, just a minor fix in the changelog

[scpeters]

it looks fine to me. the windows build failure is unrelated