README.md 2.06 KB
Newer Older
Radim Tylecek's avatar
Radim Tylecek committed
1 2 3 4 5 6 7 8 9
# 3D Reconstruction meets Semantics 

## Reconstruction Challenge

Part of the workshop is a challenge on combining 3D and semantic information in complex scenes. 
To this end, a challenging outdoor dataset, captured by a robot driving through a semantically-rich garden that contains fine geometric details, will be released. 
A multi-camera rig is mounted on top of the robot, enabling the use of both stereo and motion stereo information. 
Precise ground truth for the 3D structure of the garden has been obtained with a laser scanner and accurate pose estimates for the robot are available as well. 
Ground truth semantic labels and ground truth depth from a laser scan will be used for benchmarking the quality of the 3D reconstructions, the quality of semantic segmentation, 
Radim Tylecek's avatar
Radim Tylecek committed
10 11 12 13
and the quality of semantically annotated 3D models.

## Data

Radim Tylecek's avatar
Radim Tylecek committed
14
* File `calilbration/camchain-DDDD.yaml` - camera rig calibration
Radim Tylecek's avatar
Radim Tylecek committed
15

Radim Tylecek's avatar
Radim Tylecek committed
16
### Training
Radim Tylecek's avatar
Radim Tylecek committed
17

Radim Tylecek's avatar
Radim Tylecek committed
18 19 20 21 22
| Sequence | cameras | range | frames |
| -------- | ------- | ------ | ------- | ------ |
| train_around_hedge  | cam_0, cam_2   | 100:10:260 | 34  |
| train_boxwood_row  | cam_0, cam_2   | 90:10:650 | 114  |
| train_boxwood_slope  | cam_0, cam_2   | 120:10:340 | 46 | 
Radim Tylecek's avatar
Radim Tylecek committed
23
| train_around_roses  | cam_0, cam_2   | 590:10:690 | 11 | 
Radim Tylecek's avatar
Radim Tylecek committed
24

Radim Tylecek's avatar
Radim Tylecek committed
25
* File `model_SSSS.ply` - point cloud with semantic labels (field `scalar_s`)
Radim Tylecek's avatar
Radim Tylecek committed
26 27 28 29 30 31 32 33 34
  * Subfolders `uvc_camera_cam_X`
    * Files `uvc_camera_cam_X_fXXXXX_gtr.png` - GT annotation with reduced label set IDs (indexed bitmap)
    * Files `uvc_camera_cam_X_fXXXXX_undist.png` - undistorted color image (RGB)
    * Files `uvc_camera_cam_X_fXXXXX_pos.yml` - camera pose (xyz+quat)
    * Files `uvc_camera_cam_X_fXXXXX_over.png` - overlay of annotation over greyscale image (for display)

### Testing

| Sequence | cameras | range | frames |
Radim Tylecek's avatar
Radim Tylecek committed
35
| -------- | ------- | ------ | ------- | ------ |
Radim Tylecek's avatar
Radim Tylecek committed
36
| test_around_garden  | cam_0, cam_2   | 140:10:580, 700:10:1480 | 257 | 
Radim Tylecek's avatar
Radim Tylecek committed
37

Radim Tylecek's avatar
Radim Tylecek committed
38 39 40 41
  * Subfolders `uvc_camera_cam_X`
    * Files `uvc_camera_cam_X_fXXXXX_undist.png` - undistorted color image (RGB)


Radim Tylecek's avatar
Radim Tylecek committed
42

Radim Tylecek's avatar
Radim Tylecek committed
43 44
## Submission

Radim Tylecek's avatar
Radim Tylecek committed
45
See http://trimbot2020.webhosting.rug.nl/events/3drms/challenge/