No Detail Found!

Covid-19 physical distance monitoring system using reverse homographic transformation technique

Author(s)
Wencel Jean C. Dacay University of Science and Technology of Southern Philippines, C.M. Recto Ave., Lapasan, Cagayan de Oro City, Philippines
Jodie Rey D. Fernandez University of Science and Technology of Southern Philippines, C.M. Recto Ave., Lapasan, Cagayan de Oro City, Philippines
Mark Lister V. Nalupa University of Science and Technology of Southern Philippines, C.M. Recto Ave., Lapasan, Cagayan de Oro City, Philippines
Abstract
The world has been disrupted due to an outbreak brought by coronavirus disease 2019 (COVID-19) with its deadly spread in over 180 countries worldwide. Vaccines are already available, however, physical distancing limits people's interactions giving no chance of the disease infecting other people. The use of surveillance video coupled with computer vision technology allows the implementation of physical distancing in public. It detects persons and calculates the distance between these people. In the proposed framework, it utilizes YOLO v3 to automatically detect the person in real-time. Consequently, physical distance is computed using reverse homographic projection with Euclidean distance formula. From the experimental results, YOLO v3, alongside an adjusted non-maximum suppression threshold displayed the best results in real-time detection of persons. Moreover, reverse homographic projection showed good performance in calculating the physical distance between people. To increase further the accuracy, it is recommended to try other algorithms in measuring the distance between people.
Publication Details
Page(s) 421-425
DOI DOI not available
Published Journal: Science International, Volume: 34, Issue: 5, Year: 2022

Request the full text from PASTIC

Keywords
COVID-19, Object Detection, Covid-19, object tracking, physical distancing, distance calculation
0

Citations

0

Downloads

65

Views