TL;DR; The dataset contains 1361 images supplemented with additional spatial metadata, together with 5210 annotations in a COCO-like format.
|
1361 Frames |
5210 Instance Segmentation Labels |
3 Objects Classes |
Abstract
The problem of visual pollution is a growing concern in urban areas, characterised by intrusive visual elements that can lead to overstimulation and distraction, obstructing views and causing distractions for drivers. Large-area advertising structures, such as billboards, by being effective mediums, are significant contributors to visual pollution. Illegally placed or huge billboards can also exacerbate those issues and pose safety hazards. Therefore, there is a pressing need for effective and efficient methods to identify and manage advertising structures in urban areas. This paper proposes a deep-learning-based system for automatically detecting billboards using consumer-grade unmanned aerial vehicles. Thanks to the geospatial information from the device's sensors, the position of billboards can be estimated. Side by side with the system, we share the very first dataset for billboard detection from a drone view. It contains 1361 images supplemented with spatial metadata, together with 5210 annotations.
Object's classes in dataset
Free-standing billboard
The billboard stands an individual construction, often on one or two supports.
Wall-mounted billboard
The billboard is placed on the wall of some other structure, for example, a building.
Large road sign
Additional class. The class represents road signs whose position and size are similar to billboards.
Free-standing billboard
The billboard stands an individual construction, often on one or two supports.
Wall-mounted billboard
The billboard is placed on the wall of some other structure, for example, a building.
Large road sign
Additional class. The class represents road signs whose position and size are similar to billboards.
Download dataset
Note: The dataset is licensed under the Creative Commons Attribution Non-Commercial 4.0 International.
If you want to use the dataset commercially, don't hesitate to contact the authors: vision@put.poznan.pl.
Note: The dataset is licensed under the Creative Commons Attribution Non-Commercial 4.0 International.
If you want to use the dataset commercially, don't hesitate to contact the authors: vision@put.poznan.pl.
Citation
@article{https://doi.org/10.1111/tgis.13208,
author = {Ptak, Bartosz and Kraft, Marek},
title = {Mapping urban large-area advertising structures using drone imagery and deep learning-based spatial data analysis},
journal = {Transactions in GIS},
volume = {n/a},
number = {n/a},
pages = {},
doi = {https://doi.org/10.1111/tgis.13208},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1111/tgis.13208},
eprint = {https://onlinelibrary.wiley.com/doi/pdf/10.1111/tgis.13208}
}