Visualizing simple 3D data from files on disk should be painless and portable. SimpleViz is a simple visualization library in a Docker container using Open3D which supports common file types.
- Clone this repository and
cd
into it. xhost + && COLOR=~/Documents/colorpath DEPTH=~/Documents/depthpath make viz
That's all!
Currently supported input types:
- .png, .bmp, .jpg [color]
- .npy [color, depth]
- .pfm [disparity]
If specifying a .pfm file, make sure there is a file called calib.txt
in the same folder.
The file format should follow the Middlebury Stereo format, e.g.:
cam0=[7380.523 0 1100.563; 0 7380.523 974.039; 0 0 1]
cam1=[7380.523 0 1686.196; 0 7380.523 974.039; 0 0 1]
doffs=585.633
baseline=148.452
width=2864
height=2008
ndisp=300
isint=0
vmin=27
vmax=282
dyavg=0
dymax=0
- Multiple sets of depth data, each with respective transform
- Meshes
- 'zip' mode for handling full scenes more like a blob for easier sharing