A study of nighttime vehicle detection algorithms - Open Research Data - Bridge of Knowledge

Search

A study of nighttime vehicle detection algorithms

wersja 1.0

Description

This dataset is from my master's thesis "A study of nighttime vehicle detection algorithms". It contains both raw data and preprocessed dataset ready to use. In the pictures below you can see how images were annotated.

Package contains 3 datasets for:

  • detection (YOLO, Pascal) - 2000 cars
  • classification of images - 100 000 images
  • classification of light intensity samples - 14 000 samples

Detection dataset has two classes: "front" and "rear".

Classification datasets have two classes: "There is a vehicle" or "There is no vehicle".

Dataset designed for vehicle detection at nighttime conditions, including images and recordings from the light intensity sensor. The task of vehicle detection at night is challenging due to limited visibility and reduced image quality. The Go Pro 5 camera and APDS9966 light intensity sensor, both of which are commercially available and popular for light recording applications, were used to create this dataset.

The choice of the Go Pro 5 camera and light intensity sensor as data sources was due to their popularity and availability in the market, which allows easy replication of the experiments and expand this dataset in the future.

Illustration of the publication
Illustration of the publication
Illustration of the publication

Two settings were used - urban and rural. The urban one is characterized by high light pollution.

Illustration of the publication

Examples of "front" class.

Illustration of the publication

Examples that are not classified as "front" class.

Illustration of the publication

Examples of "rear" class.

Illustration of the publication

Examples that are not classified as "rear" class.

Dataset file

data.zip
16.0 GB, S3 ETag bdc8eb3b079e8ad9ae63dc233280c6cc-33, downloads: 54
The file hash is calculated from the formula
hexmd5(md5(part1)+md5(part2)+...)-{parts_count} where a single part of the file is 512 MB in size.

Example script for calculation:
https://github.com/antespi/s3md5
download file data.zip

File details

License:
Creative Commons: 0 1.0 open in new tab
CC 0
Public Domain Dedication

Details

Year of publication:
2023
Related location:
Gdańsk, województwo pomorskie, Polska (54° 20′ 47″ N, 18° 39′ 10″ E)
Verification date:
2023-10-10
Creation date:
2023
Dataset language:
English
Fields of science:
  • information and communication technology (Engineering and Technology)
DOI:
DOI ID 10.34808/5p14-0c72 open in new tab
Verified by:
Gdańsk University of Technology

Keywords

Cite as

Authors

Version this document has several versions

  • Current version
    version 1.0
    DOI ID10.34808/5p14-0c72
    release date 2023-10-10
DOI 10.34808/e8ht-d443 represents the latest version of the data.

seen 233 times