Index of /hyko-dataset/HyKo2

Name	Last modified	Size
Parent Directory		-
nir/	2022-07-19 16:14	-
labels/	2017-10-22 23:04	-
calib/	2017-10-22 22:57	-
raw-data/	2017-10-17 16:25	-
vis/	2017-10-17 16:16	-
readme.txt	2017-10-17 16:16	1.9K

-------------------
GENERAL INFORMATION
-------------------

1. Title of Dataset:
HyKo2

2. Author Information
Name: Christian Winkens
Institution: University of Koblenz-Landau
Address: UniversitÃ¤tsstraÃŸe 1, 56070 Koblenz
Email: cwinkens@uni-koblenz.de

3. Date of data collection
18.11.2016

4. Geographic location of data collection (where was data collected?):
Koblenz, Germany

---------------------
DATA & FILE OVERVIEW
---------------------

1. File List
A. Folder: vis
Short description: Contains extracted data from the VIS camera as annotated MATLAB Level 5 MAT files.

B. Folder: nir
Short description: Contains extracted data from the NIR camera as annotated MATLAB Level 5 MAT files.

C. Folder: raw-data
Short description: Contains raw-data in form of ros bagfiles.

C. Folder: calib
Short description: Contains calibration information.

D. Folder: labels
Short description: Contains label files.

2. Relationship between files: The image data contained in the folders were extracted from various rosbags, which were recorded in Koblenz and the surrounding area on November 18,2016. The images were extracted from the dataset about every 4 seconds, preprocessed and then saved. The masks were created with an annotation tool developed for this purpose. The data streams of the cameras were synchronized with the help of a hardware trigger. The dataset was labelled in terms of drivability, semantic classes and spectral reflectivity.

--------------------------
METHODOLOGICAL INFORMATION
--------------------------

1. Description of methods used for collection/generation of data:
The cameras used for the recording are two special multi-spectral cameras from Ximea:
- MQ022HG-IM-SM4X4-VIS with LM5JC10M lens from Kowa
- MQ022HG-IM-SM5X5-NIR with VIS-NIR lens from Edmund Optics

The recording was synchronized via a hardware trigger consisting of an Arduino board.
The data streams were recorded with Ubuntu 14.04 and ROS-Indigo and stored as rosbags.
The Rosbags are raw data that have to be pre-processed specially.

--------------------------
Labels
--------------------------

* Drivability
* Semantic
* Material