MCPcopy Create free account
hub / github.com/google-deepmind/dm_control / render

Method render

dm_control/mujoco/engine.py:178–233  ·  view source on GitHub ↗

Returns a camera view as a NumPy array of pixel values. Args: height: Viewport height (number of pixels). Optional, defaults to 240. width: Viewport width (number of pixels). Optional, defaults to 320. camera_id: Optional camera name or index. Defaults to -1, the free

(
      self,
      height=240,
      width=320,
      camera_id=-1,
      overlays=(),
      depth=False,
      segmentation=False,
      scene_option=None,
      render_flag_overrides=None,
      scene_callback: Optional[Callable[['Physics', mujoco.MjvScene],
                                        None]] = None,
  )

Source from the content-addressed store, hash-verified

176 mujoco.mj_step(self.model.ptr, self.data.ptr, nstep)
177
178 def render(
179 self,
180 height=240,
181 width=320,
182 camera_id=-1,
183 overlays=(),
184 depth=False,
185 segmentation=False,
186 scene_option=None,
187 render_flag_overrides=None,
188 scene_callback: Optional[Callable[['Physics', mujoco.MjvScene],
189 None]] = None,
190 ):
191 """Returns a camera view as a NumPy array of pixel values.
192
193 Args:
194 height: Viewport height (number of pixels). Optional, defaults to 240.
195 width: Viewport width (number of pixels). Optional, defaults to 320.
196 camera_id: Optional camera name or index. Defaults to -1, the free
197 camera, which is always defined. A nonnegative integer or string
198 corresponds to a fixed camera, which must be defined in the model XML.
199 If `camera_id` is a string then the camera must also be named.
200 overlays: An optional sequence of `TextOverlay` instances to draw. Only
201 supported if `depth` is False.
202 depth: If `True`, this method returns a NumPy float array of depth values
203 (in meters). Defaults to `False`, which results in an RGB image.
204 segmentation: If `True`, this method returns a 2-channel NumPy int32 array
205 of label values where the pixels of each object are labeled with the
206 pair (mjModel ID, mjtObj enum object type). Background pixels are
207 labeled (-1, -1). Defaults to `False`, which returns an RGB image.
208 scene_option: An optional `wrapper.MjvOption` instance that can be used to
209 render the scene with custom visualization options. If None then the
210 default options will be used.
211 render_flag_overrides: Optional mapping specifying rendering flags to
212 override. The keys can be either lowercase strings or `mjtRndFlag` enum
213 values, and the values are the overridden flag values, e.g.
214 `{'wireframe': True}` or `{mujoco.mjtRndFlag.mjRND_WIREFRAME: True}`.
215 See `mujoco.mjtRndFlag` for the set of valid flags. Must be None if
216 either `depth` or `segmentation` is True.
217 scene_callback: Called after the scene has been created and before
218 it is rendered. Can be used to add more geoms to the scene.
219
220 Returns:
221 The rendered RGB, depth or segmentation image.
222 """
223 camera = Camera(
224 physics=self,
225 height=height,
226 width=width,
227 camera_id=camera_id,
228 scene_callback=scene_callback)
229 image = camera.render(
230 overlays=overlays, depth=depth, segmentation=segmentation,
231 scene_option=scene_option, render_flag_overrides=render_flag_overrides)
232 camera._scene.free() # pylint: disable=protected-access
233 return image
234
235 def get_state(self, sig=None):

Callers

nothing calls this directly

Calls 3

renderMethod · 0.95
CameraClass · 0.85
freeMethod · 0.45

Tested by

no test coverage detected