Abstract: At the end of December 2019, Wuhan city in China was hit by the coronavirus illness 2019 (COVID-19). At this point, the infection has spread all over. The majority of governments have adopted a variety of steps to halt the spread of the disease. To prevent the transmission of COVID-19, people should keep a safe distance from one another. SD is one of the most effective treatments. Monitor social separation using a model that incorporates the data from YOLOv4. Starting with a video or photo, the model generates warnings for SD violations. At universities, malls, railway stations, and other public places, we deployed a deep learning system called YOLOv4 to better analyse pedestrian behaviour. A warning sub-system is activated as soon as the SD threshold (SDTH) or violation index (VI) are breached, resulting in an immediate awareness action. In addition, the current literature on SD, object recognition methods, and SD monitoring is thoroughly investigated and discussed in this work. While COVID-19 is still a threat, the model offered is designed to keep track of people in the areas where it is most prevalent.