pybullet_rendering

PyBullet模拟器的外部渲染。

安装

使用点子

pip install pybullet_rendering

从源代码

安装 NumPy

pip install numpy

从源代码安装 PyBullet

git clone https://github.com/bulletphysics/bullet3.git
cd bullet3
python setup.py install
export BULLET_ROOT_DIR="$PWD"

安装包

cd ..
git clone https://github.com/ikalevatykh/pybullet_rendering.git
cd pybullet_rendering
python setup.py install --bullet_dir "$BULLET_ROOT_DIR"

运行测试

python -m unittest discover tests -v

运行示例

在 GUI 模式下使用Panda3D进行渲染的示例：

python -m pybullet_rendering.examples.panda3d_gui --multisamples 8

不同渲染器的测试性能：

python -m pybullet_rendering.examples.performance

渲染器

这个包提供了基于Panda3D和pyrender的示例渲染器。

使用示例

import pybullet as pb
from pybullet_rendering import RenderingPlugin
from pybullet_rendering.render.panda3d import P3dRenderer # panda3d-based renderer
from pybullet_rendering.render.pyrender import PyrRenderer # pyrender-based renderer

client_id = pb.connect(pb.DIRECT)

# bind your renderer to pybullet
renderer = P3dRenderer(multisamples=4) # or PyrRenderer(platform='egl', egl_device=1)
plugin = RenderingPlugin(client_id, renderer)

# render thru the standard pybullet API
w, h, rgba, depth, mask = pb.getCameraImage(w, h, projectionMatrix=..., viewMatrix=...)

在 Python 中实现自己的渲染器

您的渲染器应该从BaseRenderer该类继承并实现其update_scene和render_frame方法。要了解它们的参数，请参阅示例和测试。

from pybullet_rendering import BaseRenderer

class MyRenderer(BaseRenderer):

    def __init__(self):
        """The base class initializer have to be called first."""
        BaseRenderer.__init__(self)

    def update_scene(self, scene_graph, materials_only):
        """Update a scene using scene_graph description

        Arguments:
            scene_graph {SceneGraph} -- scene description
            materials_only {bool} -- update only shape materials
        """
        for uid, pb_node in scene_graph.nodes.items():
            "update nodes of your scene"

    def render_frame(self, scene_state, scene_view, frame):
        """Render a scene at scene_state with a scene_view settings

        Arguments:
            scene_state {SceneState} --  scene state, e.g. transformations of all objects
            scene_view {SceneView} -- view settings, e.g. camera, light, viewport parameters
            frame {FrameData} -- output image buffer
        """
        w, h = scene_view.viewport        
        if return_to_bullet:
            frame.color_img[:] = ... # np.uint8, (h,w,4)
            frame.depth_img[:] = ... # np.float32 (h,w)
            frame.mask_img[:] = ... # np.int32 (h,w)
            return True

        return False  

引文

如果您发现 pybullet_rendering 在您的研究中有用，请使用以下 BibTeX 条目引用存储库。

@Misc{kalevatykh2020pybullet_rendering,
  author =       {Kalevatykh, Igor et al.},
  title =        {pybullet_rendering - External rendering for the PyBullet simulator},
  howpublished = {Github},
  year =         {2020},
  url =          {https://github.com/ikalevatykh/pybullet_rendering}
}

执照

mano_pybullet 在GPLv3许可下发布。