Abstract | ||
---|---|---|
The ever-improving Internet of things (IoT) and computer vision technologies have enabled automated monitoring of animals, which is essential for understanding animal behavior and conservation of ecosystem. The tradeoff between survey cost and sampling variability is a significant issue in designing a camera survey considering the risk of losing informative images; the monitoring accuracy tends to decrease in accordance with data reduction. However, there has been no designing method for time-lapse realtime monitoring over networks to guarantee monitoring accuracy. To address this problem, this paper proposes a Realtime Animal Monitoring over Network (RAMNe). The goal of RAMNe is to efficiently detect target animals in realtime using network cameras. We propose a determination method for the monitoring interval to guarantee the target value of monitoring accuracy based on a formal theoretical analysis using the Age of Information (AoI). The proposed scheme can minimize the amount of transferred data to enable efficient and stable monitoring even in resource-limited environments. The performance of RAMNe was evaluated with ns-3 simulations to confirm the relationship between monitoring accuracy and interval. |
Year | DOI | Venue |
---|---|---|
2020 | 10.1109/ICCWorkshops49005.2020.9145235 | 2020 IEEE International Conference on Communications Workshops (ICC Workshops) |
Keywords | DocType | ISSN |
IoT,Age of Information,scheduling,computer vision,realtime system | Conference | 2164-7038 |
ISBN | Citations | PageRank |
978-1-7281-7440-2 | 0 | 0.34 |
References | Authors | |
0 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yu Nakayama | 1 | 16 | 15.16 |
Yoshiaki Inoue | 2 | 21 | 2.06 |
Marie Katsurai | 3 | 1 | 3.05 |