Datasets:
Tasks:
Image Segmentation
Modalities:
Image
Formats:
parquet
Sub-tasks:
semantic-segmentation
Languages:
English
Size:
10K - 100K
License:
Update README.md
Browse files
README.md
CHANGED
@@ -1,4 +1,31 @@
|
|
1 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
dataset_info:
|
3 |
features:
|
4 |
- name: color
|
@@ -23,3 +50,36 @@ configs:
|
|
23 |
- split: train
|
24 |
path: data/train-*
|
25 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
+
license: cc0-1.0
|
5 |
+
license_name: cc0-1.0
|
6 |
+
license_link: https://creativecommons.org/publicdomain/zero/1.0/
|
7 |
+
tags:
|
8 |
+
- computer-vision
|
9 |
+
- autonomous-driving
|
10 |
+
- mars
|
11 |
+
- semantic-segmentation
|
12 |
+
- robotics
|
13 |
+
- space
|
14 |
+
annotations_creators:
|
15 |
+
- crowdsourced
|
16 |
+
- expert-generated
|
17 |
+
language_creators:
|
18 |
+
- found
|
19 |
+
language_details: en-US
|
20 |
+
pretty_name: 'BASEPROD: The Bardenas SemiDesert Planetary Rover Dataset'
|
21 |
+
size_categories:
|
22 |
+
- 10K<n<100K
|
23 |
+
source_datasets:
|
24 |
+
- original
|
25 |
+
task_categories:
|
26 |
+
- image-segmentation
|
27 |
+
task_ids:
|
28 |
+
- semantic-segmentation
|
29 |
dataset_info:
|
30 |
features:
|
31 |
- name: color
|
|
|
50 |
- split: train
|
51 |
path: data/train-*
|
52 |
---
|
53 |
+
|
54 |
+
# BASEPROD: The Bardenas SemiDesert Planetary Rover Dataset
|
55 |
+
|
56 |
+
## Dataset Summary
|
57 |
+
BASEPROD is a planetary rover dataset collected in the Bardenas semi-desert in Spain, containing approximately 36,000 synchronized sets of RGB, depth, and thermal images from a Realsense camera and thermal sensor, plus 62,000 additional stereo pairs from a Bumblebee XB3 camera. The dataset was collected using the MaRTA rover (Martian Rover Testbed for Autonomy) developed by ESA, traversing approximately 1.7km of Mars-analog terrain.
|
58 |
+
|
59 |
+
This is merely a reupload of the existing dataset. For more information about the dataset and its creation, please refer to [ESA Robotics Datasets](https://roboshare.esa.int/datasets/index.php/baseprod/)
|
60 |
+
|
61 |
+
## Supported Tasks
|
62 |
+
|
63 |
+
The dataset can be used for, among others:
|
64 |
+
* Image Segmentation
|
65 |
+
* Terrain Classification
|
66 |
+
* Multi-modal Vision Analysis
|
67 |
+
|
68 |
+
## Sensors
|
69 |
+
|
70 |
+
Realsense D435i RGB-D camera:
|
71 |
+
* RGB resolution: 1280x720 pixels.
|
72 |
+
* Depth resolution: 848x480 pixels.
|
73 |
+
* Mounted in front center, tilted down 20°.
|
74 |
+
|
75 |
+
Optris PI 640i thermal camera:
|
76 |
+
* Resolution: 640x480 pixels.
|
77 |
+
* Mounted below Realsense camera.
|
78 |
+
|
79 |
+
|
80 |
+
## Data Structure
|
81 |
+
* color: RGB images, PNG format (1280x720).
|
82 |
+
* depth: Values in millimeters.
|
83 |
+
* depth_16bit: 16-bit unsigned grayscale PNG (848x480).
|
84 |
+
* thermal: PNG format (640x480) with absolute temperature values (10-50°C).
|
85 |
+
* thermal_rgb: PNG format (640x480) with per-image normalization for enhanced local contrast.
|