Logical camera pose rotates user camera and heightmap

[osrf-migration]

Original report (archived issue) by Louise Poubel (Bitbucket: chapulina, GitHub: chapulina).

---

Try this minimal example world:

<?xml version="1.0" ?>
<sdf version="1.5">
  <world name="default">
    <include>
      <uri>model://vrc_driving_terrain</uri>
    </include>

    <include>
      <uri>model://sun</uri>
    </include>

    <include>
      <pose>0 10 0 0 0 0</pose>
      <uri>model://house_1</uri>
    </include>

    <model name="box">
      <pose>0 0 0.5 0 0 0</pose>
      <link name="link">
        <collision name="collision">
          <geometry>
            <box>
              <size>1 1 1</size>
            </box>
          </geometry>
        </collision>
        <visual name="visual">
          <geometry>
            <box>
              <size>1 1 1</size>
            </box>
          </geometry>
        </visual>

        <sensor name="logical_camera" type="logical_camera">

          <!-- try with and without this pose element -->
          <pose>0 0 0 0 0 1.57</pose>

          <logical_camera>
            <near>0.55</near>
            <far>5</far>
            <horizontal_fov>1.04719755</horizontal_fov>
            <aspect_ratio>1.778</aspect_ratio>
          </logical_camera>

          <visualize>true</visualize>
          <always_on>true</always_on>
          <update_rate>10</update_rate>
        </sensor>
      </link>
    </model>

  </world>
</sdf>

Without the sensor pose element, the world looks like this:

• Camera looking towards world's +X axis
• House on world's +Y axis
• Heightmap's "open end" towards world's +X
• User camera starting view has +X axis towards bottom right corner

With the sensor pose element (just adding that, the rest of the world is all the same!), it looks like this:

• Camera is now looking towards world's +Y axis - good, that's what we wanted
• House on world's +Y axis - ok, that hasn't changed
• Heightmap's "open end" towards world's -Y - how did that happen?!
• User camera starting view has +X axis towards top right corner - so it was rotated on the opposite direction of the logical camera

[osrf-migration]

Original comment by Steve Peters (Bitbucket: Steven Peters, GitHub: scpeters).

---

this is so weird

[osrf-migration]

Original comment by Louise Poubel (Bitbucket: chapulina, GitHub: chapulina).

---

Related to #2341

[osrf-migration]

Original comment by Shane Loretz (Bitbucket: Shane Loretz, GitHub: sloretz).

---

Seems likely that this either a rendering or GUI issue. The pose of a model seen by a logical camera seems is unaffected.

I tried a world with a logical camera and a box with 3 cases:

1. No rotation
2. 0.5 radians yaw rotation only in the<sensor>
3. 0.5 radians yaw rotation only in the <model> that has the <sensor>

In each I clicked the button for orthographic perspective then top down perspective.

The client window always aligned with the grid. When the rotation was in the <sensor> element everything appears to have been rotated in the client. However, the world pose of the box is identical in cases 1, 2, and 3. The logical camera pose is identical in case 2 and 3. It seems like physics is working with the right poses. The worlds appear identical in ode and bullet.

[osrf-migration]

Original comment by Louise Poubel (Bitbucket: chapulina, GitHub: chapulina).

---

There is a related question on answers:

http://answers.gazebosim.org/question/17876/logical-camera-refence-frame-affecting-gui-reference-frame/