VPR-DB-Lite is a tool for creating optimally sized databases (containing the minimum number of frames covering the scene) for place recognition task from RGB-D data using dominating set algo.
To use the tool, your data must be in a specific format.
- Color images in any format.
- Depth images corresponding to color images in 16-bit grayscale format.
- The trajectory containing one pose in each line in
timestamp tx ty tz qx qy qz qw
format.
Therefore, the structure of the dataset should look like this:
Example dataset
├── color
| ├── 001.png
| ├── 002.png
| ├── ...
├── depth
| ├── 001.pcd
| ├── 002.pcd
| ├── ...
└── CameraTrajectory.txt
The number of color images, depth images and poses in the trajectory file must be the same.
Please, check pipeline.ipynb
with usage example.