The datasets can be downloaded here.
For all experiments we used a Raytrix R12 color 3D-light-field-camera, with a MLA of F/2.4 aperture.
The camera is in Galilean internal configuration.
The mounted lens is a Nikon AF Nikkor F/1.8D with a 50 mm focal length.
The MLA organization is hexagonal row-aligned, and composed of 176 x 152 (width x height) micro-lenses with I = 3 different types.
The sensor is a Basler beA4000-62KC with a pixel size of s = 0.0055 mm.
The raw image resolution is 4080 x 3068.
We introduce three new datasets for three different focus distance configurations h, namely :
- R12-A for
h = 450 mm, - R12-B for
h = 1000 mm, - and R12-C for
h = inf.
Note that when changing the focus setting, the main lens moves with respect to the block MLA-sensor.
Each dataset is composed of:
- white raw plenoptic images acquired at different apertures (
N in {4, 5.66, 8, 11.31, 16}) using a light diffuser mounted on the main objective for pre-calibration step, - free-hand calibration targets acquired at various poses (in distance and orientation), separated into two subsets, one for the calibration process (16 images) and the other for reprojection error evaluation (15 images),
- a white raw plenoptic image acquired in the same luminosity condition and with the same aperture as in the calibration targets acquisition for devignetting,
- and calibration targets acquired with a controlled translation motion for quantitative evaluation, along with the depth maps computed by the Raytrix software (RxLive v4.0.50.2).
We use a 9 x 5 of 10 mm side checkerboard for R12-A,
a 8 x 5 of 20 mm for R12-B,
and a 6 x 4 of 30 mm for R12-C.
All images has been acquired using the free software MultiCam Studio (v6.15.1.3573) of the company Euresys.
The shutter speed has been set to 5 ms.
While taking white images for the pre-calibration step, the gain has been set to its maximum value.
For Raytrix data, we use their proprietary software RxLive (v4.0.50.2) to calibrate the camera, and compute the depth maps used in the evaluation.
The dataset can be downloaded here.
For all experiments we used a Raytrix R12 color 3D-light-field-camera, with a MLA of F/2.4 aperture.
The camera is in Galilean internal configuration.
The mounted lens is a a Nikon AF DC-Nikkor F/2D with a 135 mm focal length.
The MLA organization is hexagonal row-aligned, and composed of 176 x 152 (width x height) micro-lenses with I = 3 different types.
The sensor is a Basler beA4000-62KC with a pixel size of s = 0.0055 mm.
The raw image resolution is 4080 x 3068.
The dataset is capture at focus distance configuration h = 1500 mm.
The dataset is composed of:
- white raw plenoptic images acquired at different apertures (
N in {4, 5.66, 8, 11.31, 16}) using a light diffuser mounted on the main objective for pre-calibration step, - free-hand calibration targets acquired at various poses (in distance and orientation), separated into two subsets, one for the calibration process (16 images) and the other for reprojection error evaluation (15 images),
- a white raw plenoptic image acquired in the same luminosity condition and with the same aperture as in the calibration targets acquisition for devignetting,
- and calibration targets acquired with a controlled translation motion for quantitative evaluation, along with the depth maps computed by the Raytrix software (RxLive v4.0.50.2).
We use a 5 x 3 of 20 mm side checkerboard.
All images has been acquired using the free software MultiCam Studio (v6.15.1.3573) of the company Euresys.
The shutter speed has been set to 5 ms.
While taking white images for the pre-calibration step, the gain has been set to its maximum value.
For Raytrix data, we use their proprietary software RxLive (v4.0.50.2) to calibrate the camera, and compute the depth maps used in the evaluation.
Simulated dataset for Lytro-like plenoptic camera configuration, i.e., unfocused plenoptic camera (UPC).
The dataset can be downloaded here.
We used the Lytro Illum intrinsic parameters reported in Table 4 of Bok et al. (2017) as baseline for the simulation, corresponding to a main lens of aperture F/2 with a 9.9845 mm focal length.
The camera is in unfocused internal configuration (i.e., f = d).
The MLA organization is hexagonal row-aligned, and composed of 541 x 434 (width x height) micro-lenses of the same type (I = 1).
The raw image resolution is 7728 x 5368 pixel, with a pixel size of s = 0.0014 mm and with micro-image of radius 7.172 pixel.
The dataset is correspond to the focus distance configuration h = hyperfocal.
The dataset is composed of:
- white raw plenoptic images simulated at different apertures (
N in {2, 4, 5.6}) for pre-calibration step, - free-hand calibration targets simulated at various poses (in distance and orientation), separated into two subsets, one for the calibration process and the other for reprojection error evaluation,
- and calibration targets with known translation along the z-axis for quantitative evaluation.
We use a 9 x 5 of 26.25 mm side checkerboard.
All images has been generated using the libpleno and our raytracing simulator PRISM.
Datasets containing ground truth data on 3D complex real-world scene acquired with a 3D lidar scanner Leica ScanStation P20 (LP20).
The datasets can be downloaded here.
For our experiments we used a Raytrix R12 color 3D-light-field-camera, with a MLA of F/2.4 aperture.
The camera is in Galilean internal configuration.
The mounted lens is a Nikon AF Nikkor F/1.8D with a 50 mm focal length.
The MLA organization is hexagonal row-aligned, and composed of 176 x 152 (width x height) micro-lenses with I = 3 different types.
The sensor is a Basler beA4000-62KC with a pixel size of s = 0.0055 mm.
The raw image resolution is 4080 x 3068 pixel.
All images have been acquired using the MultiCamStudio free software (v6.15.1.3573) of the Euresys company.
We set the shutter speed to 5 ms.
3D scenes setup. We use a 3D lidar scanner, a Leica ScanStation P20 (LP20), that allows us to capture a color point cloud with high precision that can be used as ground truth data. The LP20 is configured with no HDR and with a resolution of 1.6 mm @ 10 m.
The configuration corresponds to a focus distance h = 2133 mm.
We built a calibration dataset, using a 6 x 4 of 30 mm side checkerboard, which is composed of:
- white raw plenoptic images acquired at different apertures (
N in {2.8, 4, 5.66, 8, 11.31, 16}) using a light diffuser mounted on the main objective for pre-calibration, - free-hand calibration target images acquired at various poses (in distance and orientation) for the calibration process (31 images),
- a white raw plenoptic image acquired in the same luminosity condition and with the same aperture as in the calibration targets acquisition for devignetting.
With this configuration, we created two sub-datasets:
- a simulated dataset built upon our own simulator PRISM based on raytracing to generate images (with
1500 rays/pixel) with known absolute position for quantitative evaluation, named R12-ES (15 images, from500 mmto1900 mmwith a step of100 mm); - a dataset composed of several 3D scenes with ground truth acquired with the LP20, for object distances ranging from
400 mmto1500 mm.
The latter dataset, named R12-ELP20, includes fives scenes:
- one scene for extrinsic parameters calibration, containing checker corner targets, named Calib;
- two scenes containing textured planar objects, named Plane-1 and Plane-2;
- and two more complex scenes containing various figurines, named Figurines-1 and Figurines-2.
Each scene is composed of: a colored point cloud (with spatial (x,y,z) information, color information (r,g,b), and intensity information) in format .ptx, .pts and .xyz; 3D positions of the targets in the lidar reference frame; two raw plenoptic images in rgb color and two raw plenoptic images in bayer; finally, photos and labels of the scene.
For instance, the datasets can be used with the following applications :
- COMPOTE (Calibration Of Multi-focus PlenOpTic camEra), a collection of tools to pre-calibrate and calibrate (multifocus) plenoptic cameras.
- PRISM (Plenoptic Raw Image Simulator), a collection of tools to generate and simulate raw images from (multifocus) plenoptic cameras.
- BLADE (BLur Aware Depth Estimation with a plenoptic camera), a collection of tools to estimate depth map from raw images obtained by (multifocus) plenoptic cameras.
Based on the libpleno, an open-source C++ computer-vision library for plenoptic cameras modeling and processing.
If you use our datasets, the libpleno, the COMPOTE tools, the PRISM tools or the BLADE tools in an academic context, please cite the following publication:
@inproceedings{labussiere2020blur,
title = {Blur Aware Calibration of Multi-Focus Plenoptic Camera},
author = {Labussi{\`e}re, Mathieu and Teuli{\`e}re, C{\'e}line and Bernardin, Fr{\'e}d{\'e}ric and Ait-Aider, Omar},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
pages = {2545--2554},
year = {2020}
}
or
@article{labussiere2022calibration
title = {Leveraging blur information for plenoptic camera calibration},
author = {Labussi{\`{e}}re, Mathieu and Teuli{\`{e}}re, C{\'{e}}line and Bernardin, Fr{\'{e}}d{\'{e}}ric and Ait-Aider, Omar},
doi = {10.1007/s11263-022-01582-z},
journal = {International Journal of Computer Vision},
year = {2022},
month = {may},
number = {2012},
pages = {1--23}
}
or
@article{labussiere2023blade,
title = {Blur aware metric depth estimation with multi-focus plenoptic cameras},
author = {Mathieu Labussière and Céline Teulière and Omar Ait-Aider},
keywords = {Plenoptic Camera, Multi-focus, Calibration, Defocus stereo, Relative blur, Disparity, Metric depth estimation},
doi = {https://doi.org/10.1016/j.cviu.2023.103802},
issn = {1077-3142},
journal = {Computer Vision and Image Understanding},
year = {2023},
volume = {235},
pages = {103802},
url = {https://www.sciencedirect.com/science/article/pii/S1077314223001820}
}
This work is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License. Enjoy!
Note: if download links are broken, don't hesitate to contact me!