Optix: Support mouse picking

**Original report ([archived issue](https://osrf-migration.github.io/ignition-gh-pages/#!/ignitionrobotics/ign-rendering/issues/19)) by Ian Chen (Bitbucket: [Ian Chen](https://bitbucket.org/%7Beaa6fca5-6deb-43f6-907f-971c144735dd%7D/), GitHub: [iche033](https://github.com/iche033)).**

----------------------------------------
Add a function that lets user pass in x, y coordinates in screen and returns the name / ID of selected entity in the scene

Ogre implementation:

- port gazebo's [selection buffer](https://bitbucket.org/osrf/gazebo/src/eddba39b82cd7a8d9d2e3cff7c6a43174bdd972b/gazebo/rendering/selection_buffer/?at=default) and offer a function similar to gazebo's [Scene::VisualAt](https://bitbucket.org/osrf/gazebo/src/eddba39b82cd7a8d9d2e3cff7c6a43174bdd972b/gazebo/rendering/Scene.cc?at=default&fileviewer=file-view-default#Scene.cc-988)

Optix:

- TBD

Start by reading Gazebo's referenced rendering/selection_buffer implementation and Scene::VisualAt API, then compare how this rendering library exposes engine-independent scene operations. Done means users can pass screen x and y coordinates and receive the selected entity's name or ID, with the Optix behavior clarified and covered.

Written by the indexing model from the issue text.