Addressing Distortion in Social Distance Monitoring: Calibration for Accurate Distance Calculation

The Social Distance Monitoring (SDM) system is a technology-driven solution that aims to enforce and maintain social distancing rules. During epidemics such as COVID-19, practising physical distancing is crucial to prevent the transmission of infectious diseases. This study focuses on the utilisation of the OpenPose algorithm for human detection, which effectively detects humans and identifies key points such as eyes-to-feet points as human height. These points enable the calculation of the distance between individuals and the width of each person. To ensure accurate distance measurement, a perspective transformation is applied to each frame after human detection, effectively addressing distortion issues caused by variations in camera angles. A novel calibration process is also introduced to handle situations where the video’s viewing angle exceeds a certain threshold, resulting in distortion. The calibration process calculates the distance ratio using frame partitioning, improving the precision of distance measurement. This study contributes to the advancement of social distance monitoring systems, providing an effective calibration solution for enforcing social distancing rules.