Low Latency and Non-Intrusive Accurate Object Detection in Forests

Hussain, Ambreen and Barua, Bidushi and Osman, Ahmed and Abozariba, Raouf and Asyhari, A. Taufiq (2022) Low Latency and Non-Intrusive Accurate Object Detection in Forests. In: IEEE Symposium Series on Computational Intelligence (IEEE SSCI 2021), 5-7 December 2021, Orlando, Florida, USA.

Preview

Text 2021217326.pdf - Accepted Version Download (2MB)

Abstract

A resilient and healthy forest requires accurate and timely monitoring, observing key forest health indicators (FHI). Forest managers and rangers usually perform tedious manual data collection using citizen science for biodiversity conservation and ecological research. With the advent of faster radio network technologies such as 4G, it is advantageous to leverage these networks’ high speed and low latency for real-time monitoring. We present a novel approach to stream high definition videos over cellular networks to provide real-time (< 0.5 seconds) data transmission to the YOLOv5 machine learning algorithm, hosted in the cloud. The system provides non-intrusive precise tree class detection, matching existing models such as Fast R-CNN and SSD. Our investigation also reveals that the same accuracy can be achieved with 99% fewer iterations, minimizing computational time and cost.

Item Type:	Conference or Workshop Item (Paper)
Identification Number:	10.1109/SSCI50451.2021.9660175
Dates:	Date Event 2 October 2021 Accepted 24 January 2022 Published Online
Uncontrolled Keywords:	YOLOv5, 4G/5G, WebRTC, tree species detection, low-latency
Subjects:	CAH11 - computing > CAH11-01 - computing > CAH11-01-04 - software engineering CAH11 - computing > CAH11-01 - computing > CAH11-01-05 - artificial intelligence
Divisions:	Faculty of Computing, Engineering and the Built Environment > College of Computing
Depositing User:	Ambreen Hussain
Date Deposited:	29 Nov 2021 13:10
Last Modified:	22 Mar 2023 12:00
URI:	https://www.open-access.bcu.ac.uk/id/eprint/12482

Actions (login required)

View Item