https://github.com/OuyangJunyuan/PointCloudRenderer
a point cloud renderer based on mitsuba3
https://github.com/OuyangJunyuan/PointCloudRenderer
pointcloud rendering
Last synced: about 1 year ago
JSON representation
a point cloud renderer based on mitsuba3
- Host: GitHub
- URL: https://github.com/OuyangJunyuan/PointCloudRenderer
- Owner: OuyangJunyuan
- Created: 2022-07-03T11:58:56.000Z (almost 4 years ago)
- Default Branch: main
- Last Pushed: 2023-03-10T05:55:24.000Z (about 3 years ago)
- Last Synced: 2024-08-01T03:39:11.228Z (almost 2 years ago)
- Topics: pointcloud, rendering
- Language: Python
- Homepage:
- Size: 21 MB
- Stars: 33
- Watchers: 2
- Forks: 3
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Point Cloud Renderer
# Update
* 2023-03-10:
* Currently, it is compatible with mitsuba3.
* Support GIF recording for moving views or points.
# Preparation
`pip install -r requirements.txt`
## Preliminary
* point cloud format： the extension is `npy` (saved by numpy ), storing a list of [x,y,z,...].
* coordinate definition:
* left: x
* up:z
* screen facing out: y
[//]: # (漫反射)
[//]: # ()
[//]: # ()
[//]: # ()
[//]: # (透明)
[//]: # ()
[//]: # (反光)
[//]: # ()
## single image rendering
```shell
python scripts/once.py \
-f workspace/demo/car.npy \
-o ./workspace/demo/results/car_single \
--format xyzi \
--color_by i --color_normalize \
--pose 0 0 0 0 0 -180
--preview
```
arguments:
* -f: path to input point cloud
* -o: path to output image
* --format: the format of the given point cloud, e.g., xyz/xyzi/xyzrgbs/...
* --color_by: which data field is used to render colors when the point cloud file does not provide rgb field for each point.
* --color_normalize: normalize color_by
* --pose: the pose [x y z r p y] of the given points (in degrees).
* --preview: for quickly rendering.
besides, you can modify the config file `config.yaml` to adjust the size of images or other parameters of the renderer:
* sample: sample per pixel, the higher, the better rendering quality.
## a rotating object
```shell
python scripts/anim_obj_axis.py \
-f workspace/demo/car.npy \
-o ./workspace/demo/results/car_rot_z \
--pose 0 0 0 0 0 -180 \
--axis 3 0 360 5 \
--fps 15
```
arguments:
* -axis: [axis order(x:1,y:2,z:3), begin, end ,interval]
* -fps: the frames per second for gif animation.
## given pose list
```shell
python scripts/anim_obj.py \
-f workspace/demo/car.npy \
-o ./workspace/demo/results/car_rot_z \
--pose 0 0 0 0 0 -180 \
--pose_file ./workspace/demo/poses.txt \
--fps 15
```
## traveling in scene
```shell
python scripts/anim_view.py \
-f workspace/demo/car.npy \
-o ./workspace/demo/results/car_views \
--pose 0 0 0 0 0 -180 \
--view_file ./workspace/demo/views.txt \
--fps 2
```
arguments:
* --view_file: store a list of (camera_x,camera_y,camera_z,target_x,targe_y,targe_z)
# Further development
1. The official reference link is given inline the code about scene definition.
2. Edit the scene: [ref](git@github.com:OuyangJunyuan/PointCloudRenderer.git)
3. for more properties of plugins please refer
this [link](https://mitsuba.readthedocs.io/en/stable/src/rendering_tutorials.html).
# Acknowledge
This script deeply based on[Mitsuba2PointCloudRenderer](https://github.com/tolgabirdal/Mitsuba2PointCloudRenderer) and [PointFlowRenderer](https://github.com/zekunhao1995/PointFlowRenderer) and is an easy-to-use version.