Object Detection in Unmanned Aerial Vehicle Camera Stream Using Deep Neural Network

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26220%2F22%3APU146264" target="_blank" >RIV/00216305:26220/22:PU146264 - isvavai.cz</a>

Výsledek na webu

<a href="https://ieeexplore.ieee.org/xpl/conhome/1800005/all-proceedings" target="_blank" >https://ieeexplore.ieee.org/xpl/conhome/1800005/all-proceedings</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/ICUMT57764.2022.9943463" target="_blank" >10.1109/ICUMT57764.2022.9943463</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Object Detection in Unmanned Aerial Vehicle Camera Stream Using Deep Neural Network

Popis výsledku v původním jazyce

Nowadays, the world is experiencing an increasing boom in applications of artificial intelligence, especially deep learning. This is more and more used in many areas such as industry, medicine and security systems, etc. This article deals with object detection from Unmanned Aerial Vehicle (UAV) perspective. The whole system uses one camera, which is suitably positioned on the UAV to capture the scene. Image processing and subsequent object detection using the YOLOv4 model are performed on the Jetson Nano device. The device itself is relatively powerful, but to save the computing power of the device, the YOLOv4 neural network model was modified. The YOLOv4 model was trained on our own dataset. This training set was created specifically for UAV applications. The result of this work is a learned YOLOv4 neural network model designed for UAVs with regard to the used training set. The modified network model is also able to run in real-time and save computing power for possibly other UAV operations. All materials, dataset and scripts used in this work, are available at https://github.com/KicoSVK/object-detection-in-uav-using-yolov4.

Název v anglickém jazyce

Object Detection in Unmanned Aerial Vehicle Camera Stream Using Deep Neural Network

Popis výsledku anglicky

Nowadays, the world is experiencing an increasing boom in applications of artificial intelligence, especially deep learning. This is more and more used in many areas such as industry, medicine and security systems, etc. This article deals with object detection from Unmanned Aerial Vehicle (UAV) perspective. The whole system uses one camera, which is suitably positioned on the UAV to capture the scene. Image processing and subsequent object detection using the YOLOv4 model are performed on the Jetson Nano device. The device itself is relatively powerful, but to save the computing power of the device, the YOLOv4 neural network model was modified. The YOLOv4 model was trained on our own dataset. This training set was created specifically for UAV applications. The result of this work is a learned YOLOv4 neural network model designed for UAVs with regard to the used training set. The modified network model is also able to run in real-time and save computing power for possibly other UAV operations. All materials, dataset and scripts used in this work, are available at https://github.com/KicoSVK/object-detection-in-uav-using-yolov4.

Druh

D - Stať ve sborníku

CEP obor

—

OECD FORD obor

20205 - Automation and control systems

Projekt

<a href="/cs/project/FW01010548" target="_blank" >FW01010548: Detekce, identifikace a sledování objektů z UAV pomocí strojového vidění</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2022

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

Proceedings of the 2021 13th International Congress on Ultra Modern Telecommunications and Control Systems and Workshops (ICUMT)

ISBN

979-8-3503-9866-3

ISSN

2157-023X

e-ISSN

—

Počet stran výsledku

5

Strana od-do

80-84

Název nakladatele

IEEE

Místo vydání

Valencia, Spain

Místo konání akce

Valencia, Spain

Datum konání akce

11. 10. 2022

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

—