On-Board Crowd Counting and Density Estimation Using Low Altitude Unmanned Aerial Vehicle-Looking beyond Beating the Benchmark

Bartosz Ptak; Dominik Pieczyński; Mateusz Piechocki; Marek Kraft

doi:10.3390/rs14102288

Scientific Information System of the Poznań University of Technology

PL EN

Main page / Publications / On-Board Crowd Counting and Density Estimation Using Low Altitude Unmanned Aerial Vehicle-Looking beyond Beating the Benchmark

Submit a comment

Article

Download file Download BibTeX

Title

On-Board Crowd Counting and Density Estimation Using Low Altitude Unmanned Aerial Vehicle-Looking beyond Beating the Benchmark

Authors

Bartosz Ptak (WARiE) ^{[ 1 ][ 2.2 ][ SzD ]}
Dominik Pieczyński (WARiE) ^{[ 2 ][ 2.2 ][ P ]}
Mateusz Piechocki (WARiE) ^{[ 1 ][ 2.2 ][ SzD ]}
Marek Kraft (WARiE) ^{[ 2 ][ 2.2 ][ P ]}

^{[ 1 ]} Wydział Automatyki, Robotyki i Elektrotechniki, Politechnika Poznańska | ^{[ 2 ]} Instytut Robotyki i Inteligencji Maszynowej, Wydział Automatyki, Robotyki i Elektrotechniki, Politechnika Poznańska | ^{[ SzD ]} doctoral school student | ^{[ P ]} employee

Scientific discipline (Law 2.0)

[2.2] Automation, electronics, electrical engineering and space technologies

Year of publication

2022

Published in

Remote Sensing

Journal year: 2022 | Journal volume: vol. 14 | Journal number: iss. 10

Article type

scientific article

Publication language

english

Keywords

EN

deep learning
crowd counting
image processing
edge processing
embedded
UAV

Abstract

EN Recent advances in deep learning-based image processing have enabled significant improvements in multiple computer vision fields, with crowd counting being no exception. Crowd counting is still attracting research interest due to its potential usefulness for traffic and pedestrian stream monitoring and analysis. This study considered a specific case of crowd counting, namely, counting based on low-altitude aerial images collected by an unmanned aerial vehicle. We evaluated a range of neural network architectures to find ones appropriate for on-board image processing using edge computing devices while minimising the loss in performance. Through experiments on a range of neural network architectures, we also showed that the input image resolution significantly impacts the prediction quality and should be considered an important factor before going for a more complex neural network model to improve accuracy. Moreover, by extending a state-of-the-art benchmark with more in-depth testing, we showed that larger models might be prone to overfitting because of the relative scarcity of training data.

Pages (from - to)

2288-1 - 2288-18

DOI

10.3390/rs14102288

URL

https://www.mdpi.com/2072-4292/14/10/2288/htm

Comments

Article number: 2288

License type

CC BY (attribution alone)

Open Access Mode

open journal