Method render

dm_control/mujoco/engine.py:840–946

Renders the camera view as a numpy array of pixel values. Args: overlays: An optional sequence of `TextOverlay` instances to draw. Only supported if `depth` and `segmentation` are both False. depth: An optional boolean. If True, makes the camera return depth measurem

(
      self,
      overlays=(),
      depth=False,
      segmentation=False,
      scene_option=None,
      render_flag_overrides=None,
  )

                          self._physics.contexts.mujoco.ptr)

  def render(
      self,
      overlays=(),
      depth=False,
      segmentation=False,
      scene_option=None,
      render_flag_overrides=None,
  ):
    """Renders the camera view as a numpy array of pixel values.

    Args:
      overlays: An optional sequence of `TextOverlay` instances to draw. Only
        supported if `depth` and `segmentation` are both False.
      depth: An optional boolean. If True, makes the camera return depth
        measurements. Cannot be enabled if `segmentation` is True.
      segmentation: An optional boolean. If True, make the camera return a
        pixel-wise segmentation of the scene. Cannot be enabled if `depth` is
        True.
      scene_option: A custom `wrapper.MjvOption` instance to use to render
        the scene instead of the default.  If None, will use the default.
      render_flag_overrides: Optional mapping containing rendering flags to
        override. The keys can be either lowercase strings or `mjtRndFlag` enum
        values, and the values are the overridden flag values, e.g.
        `{'wireframe': True}` or `{mujoco.mjtRndFlag.mjRND_WIREFRAME: True}`.
        See `mujoco.mjtRndFlag` for the set of valid flags. Must be empty if
        either `depth` or `segmentation` is True.

    Returns:
      The rendered scene.
        * If `depth` and `segmentation` are both False (default), this is a
          (height, width, 3) uint8 numpy array containing RGB values.
        * If `depth` is True, this is a (height, width) float32 numpy array
          containing depth values (in meters).
        * If `segmentation` is True, this is a (height, width, 2) int32 numpy
          array where the first channel contains the integer ID of the object at
          each pixel, and the second channel contains the corresponding object
          type (a value in the `mjtObj` enum). Background pixels are labeled
          (-1, -1).

    Raises:
      ValueError: If either `overlays` or `render_flag_overrides` is requested
        when `depth` or `segmentation` rendering is enabled.
      ValueError: If both depth and segmentation flags are set together.
    """

    if overlays and (depth or segmentation):
      raise ValueError(_OVERLAYS_NOT_SUPPORTED_FOR_DEPTH_OR_SEGMENTATION)

    if render_flag_overrides and (depth or segmentation):
      raise ValueError(
          _RENDER_FLAG_OVERRIDES_NOT_SUPPORTED_FOR_DEPTH_OR_SEGMENTATION)

    if depth and segmentation:
      raise ValueError(_BOTH_SEGMENTATION_AND_DEPTH_ENABLED)

    if render_flag_overrides is None:
      render_flag_overrides = {}