search

NVlabs / dex-ycb-toolkit

A Python package that provides evaluation and visualization tools for the DexYCB dataset
https://dex-ycb.github.io
GNU General Public License v3.0
145 stars 24 forks source link

Camera parameters #1

Open anilesec opened 3 years ago

anilesec commented 3 years ago

Thanks for the really cool work and the dataset. I am just wondering, does that released dataset contains camera parameters(extrinsic and intrinsic)?

Thank you!

ychao-nvidia commented 3 years ago

Yes, it does. The intrinsics and extrinsics are provided in the dataset (see the project site for download).

anilesec commented 2 years ago

I checked the calibration folder, but there are 10 extrinsic.yml files. It is supposed to be 8, right? because we only have 8 fixed cameras. Additionally, each extrinsic.yml file has more than one set of extrinsics. I do not understand how to decode this yml file. Could you please explain how to understand the info in extrinsic.yml file?

Thank you!

ychao-nvidia commented 2 years ago

The 10 extrinsic.yml files correspond to the 10 subjects. We captured each subject in different dates so we re-calibrated the cameras before capturing a subject. If you open meta.yml in a sequence's folder, you'll see the extrinsics ID for that sequence, e.g., 20200702_151821 for the sequence 20200709-subject-01/20200709_141754.

Each extrinsics.yml file contains the extrinsics for all the 8 cameras. For example, if you open extrinsics_20200702_151821/extrinsics.yml, you should see:

extrinsics:
  '836212060125': !!python/tuple
  - -0.894641101360321
  - -0.04048347473144531
  - -0.44494757056236267
  - 0.6362327337265015
  - -0.31882184743881226
  - 0.7555275559425354
  - 0.5723029971122742
  - -0.6008626222610474
  - 0.3130013048648834
  - 0.6538649201393127
  - -0.6888331174850464
  - 1.3911463022232056
  '839512060362': !!python/tuple
  - -0.22983123362064362
  - 0.262200266122818
  - 0.9372453093528748
  - -0.6329705715179443
  - 0.7641939520835876
  - 0.6449487209320068
  - 0.006967182271182537
  - -0.07978752255439758
  - -0.6026484370231628
  - 0.71783846616745
  - -0.34860125184059143
  - 1.0656930208206177
...

836212060125 is the ID of one camera and the following numbers represent its extrinsics in a 3x4 transformation matrix (for rotation and translation).

To understand how we consume these extrinsics files, you can look at the example of the interactive 3D viewer. The extrinsics are loaded in this line.