Rendering¶

Classes for rendering images of meshes.

VirtualCamera¶

class meshpy.VirtualCamera(camera_intr)¶

Bases: object

A virtualized camera for rendering virtual color and depth images of meshes.

Rendering is performed by using OSMesa offscreen rendering and boost_numpy.

__init__(camera_intr)¶

Initialize a virtual camera.

Parameters:	camera_intr (`CameraIntrinsics`) – The CameraIntrinsics object used to parametrize the virtual camera.
Raises:	`ValueError` – When camera_intr is not a CameraIntrinsics object.

add_to_scene(name, scene_object)¶

Add an object to the scene.

Parameters:	name (`str`) – name of object in the scene scene_object (`SceneObject`) – object to add to the scene

remove_from_scene(name)¶

Remove an object to a from the scene.

Parameters:	name (`str`) – name of object to remove

images(mesh, object_to_camera_poses, mat_props=None, light_props=None, enable_lighting=True, debug=False)¶

Render images of the given mesh at the list of object to camera poses.

Parameters:	mesh (`Mesh3D`) – The mesh to be rendered. object_to_camera_poses (`list` of `RigidTransform`) – A list of object to camera transforms to render from. mat_props (`MaterialProperties`) – Material properties for the mesh light_props (`MaterialProperties`) – Lighting properties for the scene enable_lighting (bool) – Whether or not to enable lighting debug (bool) – Whether or not to debug the C++ meshrendering code.
Returns:	A 2-tuple of ndarrays. The first, which represents the color image, contains ints (0 to 255) and is of shape (height, width, 3). Each pixel is a 3-ndarray (red, green, blue) associated with a given y and x value. The second, which represents the depth image, contains floats and is of shape (height, width). Each pixel is a single float that represents the depth of the image.
Return type:	`tuple` of numpy.ndarray

images_viewsphere(mesh, vs_disc, mat_props=None, light_props=None)¶

Render images of the given mesh around a view sphere.

Parameters:	mesh (`Mesh3D`) – The mesh to be rendered. vs_disc (`ViewsphereDiscretizer`) – A discrete viewsphere from which we draw object to camera transforms. mat_props (`MaterialProperties`) – Material properties for the mesh light_props (`MaterialProperties`) – Lighting properties for the scene
Returns:	A 2-tuple of ndarrays. The first, which represents the color image, contains ints (0 to 255) and is of shape (height, width, 3). Each pixel is a 3-ndarray (red, green, blue) associated with a given y and x value. The second, which represents the depth image, contains floats and is of shape (height, width). Each pixel is a single float that represents the depth of the image.
Return type:	`tuple` of numpy.ndarray

wrapped_images(mesh, object_to_camera_poses, render_mode, stable_pose=None, mat_props=None, light_props=None, debug=False)¶

Create ObjectRender objects of the given mesh at the list of object to camera poses.

Parameters:	mesh (`Mesh3D`) – The mesh to be rendered. object_to_camera_poses (`list` of `RigidTransform`) – A list of object to camera transforms to render from. render_mode (int) – One of RenderMode.COLOR, RenderMode.DEPTH, or RenderMode.SCALED_DEPTH. stable_pose (`StablePose`) – A stable pose to render the object in. mat_props (`MaterialProperties`) – Material properties for the mesh light_props (`MaterialProperties`) – Lighting properties for the scene debug (bool) – Whether or not to debug the C++ meshrendering code.
Returns:	A list of ObjectRender objects generated from the given parameters.
Return type:	`list` of `ObjectRender`

wrapped_images_viewsphere(mesh, vs_disc, render_mode, stable_pose=None, mat_props=None, light_props=None)¶

Create ObjectRender objects of the given mesh around a viewsphere.

Parameters:	mesh (`Mesh3D`) – The mesh to be rendered. vs_disc (`ViewsphereDiscretizer`) – A discrete viewsphere from which we draw object to camera transforms. render_mode (int) – One of RenderMode.COLOR, RenderMode.DEPTH, or RenderMode.SCALED_DEPTH. stable_pose (`StablePose`) – A stable pose to render the object in. mat_props (`MaterialProperties`) – Material properties for the mesh light_props (`MaterialProperties`) – Lighting properties for the scene
Returns:	A list of ObjectRender objects generated from the given parameters.
Return type:	`list` of `ObjectRender`

wrapped_images_planar_worksurface(mesh, ws_disc, render_mode, stable_pose=None, mat_props=None, light_props=None)¶

Create ObjectRender objects of the given mesh around a viewsphere and a planar worksurface, where translated objects project into the center of the camera.

Parameters:

mesh (Mesh3D) – The mesh to be rendered.
ws_disc (PlanarWorksurfaceDiscretizer) – A discrete viewsphere and translations in plane from which we draw object to camera transforms.
render_mode (int) – One of RenderMode.COLOR, RenderMode.DEPTH, or RenderMode.SCALED_DEPTH.
stable_pose (StablePose) – A stable pose to render the object in.
mat_props (MaterialProperties) – Material properties for the mesh
light_props (MaterialProperties) – Lighting properties for the scene

Returns:

list of ObjectRender – A list of ObjectRender objects generated from the given parameters.
list of RigidTransform – A list of the transformations from object frame to camera frame used for rendering the images.
list of CameraIntrinsics – A list of the camera intrinsics used for rendering the images.

ViewsphereDiscretizer¶

class meshpy.ViewsphereDiscretizer(min_radius, max_radius, num_radii, min_elev, max_elev, num_elev, min_az=0, max_az=6.283185307179586, num_az=1, min_roll=0, max_roll=6.283185307179586, num_roll=1)¶

Bases: object

Set of parameters for automatically rendering a set of images from virtual cameras placed around a viewing sphere.

The view sphere indicates camera poses relative to the object.

min_radius¶: float – Minimum radius for viewing sphere.

max_radius¶: float – Maximum radius for viewing sphere.

num_radii¶: int – Number of radii between min_radius and max_radius.

min_elev¶: float – Minimum elevation (angle from z-axis) for camera position.

max_elev¶: float – Maximum elevation for camera position.

num_elev¶: int – Number of discrete elevations.

min_az¶: float – Minimum azimuth (angle from x-axis) for camera position.

max_az¶: float – Maximum azimuth for camera position.

num_az¶: int – Number of discrete azimuth locations.

min_roll¶: float – Minimum roll (rotation of camera about axis generated by azimuth and elevation) for camera.

max_roll¶: float – Maximum roll for camera.

num_roll¶: int – Number of discrete rolls.

__init__(min_radius, max_radius, num_radii, min_elev, max_elev, num_elev, min_az=0, max_az=6.283185307179586, num_az=1, min_roll=0, max_roll=6.283185307179586, num_roll=1)¶

Initialize a ViewsphereDiscretizer.

Parameters:

min_radius (float) – Minimum radius for viewing sphere.
max_radius (float) – Maximum radius for viewing sphere.
num_radii (int) – Number of radii between min_radius and max_radius.
min_elev (float) – Minimum elevation (angle from z-axis) for camera position.
max_elev (float) – Maximum elevation for camera position.
num_elev (int) – Number of discrete elevations.
min_az (float) – Minimum azimuth (angle from x-axis) for camera position.
max_az (float) – Maximum azimuth for camera position.
num_az (int) – Number of discrete azimuth locations.
min_roll (float) – Minimum roll (rotation of camera about axis generated by azimuth and elevation) for camera.
max_roll (float) – Maximum roll for camera.
num_roll (int) – Number of discrete rolls.

object_to_camera_poses()¶

Turn the params into a set of object to camera transformations.

Returns:	A list of rigid transformations that transform from object space to camera space.
Return type:	`list` of `RigidTransform`

PlanarWorksurfaceDiscretizer¶

class meshpy.PlanarWorksurfaceDiscretizer(min_radius, max_radius, num_radii, min_elev, max_elev, num_elev, min_az=0, max_az=6.283185307179586, num_az=1, min_roll=0, max_roll=6.283185307179586, num_roll=1, min_x=0, max_x=1, num_x=1, min_y=0, max_y=1, num_y=1)¶

Bases: object

Set of parameters for automatically rendering a set of images from virtual cameras placed around a viewsphere and translated along a planar worksurface.

The view sphere indicates camera poses relative to the object.

__init__(min_radius, max_radius, num_radii, min_elev, max_elev, num_elev, min_az=0, max_az=6.283185307179586, num_az=1, min_roll=0, max_roll=6.283185307179586, num_roll=1, min_x=0, max_x=1, num_x=1, min_y=0, max_y=1, num_y=1)¶

Initialize a ViewsphereDiscretizer.

Parameters:

min_radius (float) – Minimum radius for viewing sphere.
max_radius (float) – Maximum radius for viewing sphere.
num_radii (int) – Number of radii between min_radius and max_radius.
min_elev (float) – Minimum elevation (angle from z-axis) for camera position.
max_elev (float) – Maximum elevation for camera position.
num_elev (int) – Number of discrete elevations.
min_az (float) – Minimum azimuth (angle from x-axis) for camera position.
max_az (float) – Maximum azimuth for camera position.
num_az (int) – Number of discrete azimuth locations.
min_roll (float) – Minimum roll (rotation of camera about axis generated by azimuth and elevation) for camera.
max_roll (float) – Maximum roll for camera.
num_roll (int) – Number of discrete rolls.
min_x (float) – Minimum x value.
max_x (float) – Maximum x value.
num_x (int) – Number of points along x axis.
min_y (float) – Minimum y value.
may_y (float) – Mayimum y value.
num_y (int) – Number of points along y axis.

object_to_camera_poses(camera_intr)¶

Turn the params into a set of object to camera transformations.

Returns:

list of RigidTransform – A list of rigid transformations that transform from object space to camera space.
list of RigidTransform – A list of rigid transformations that transform from object space to camera space without the translation in the plane
list of CameraIntrinsics – A list of camera intrinsics that project the translated object into the center pixel of the camera, simulating cropping