Anan Nugroho, Faizal Indaryanto, Alfa Faridh Suni
The current Covid-19 pandemic has had quite an impact on society. The scale of the spread of Covid-19 is very fast, so it requires proper handling. One way to reduce the spread of Covid-19 is to practice social distancing. However, people tend to be negligent in implementing these health protocols. One way to overcome this problem is with a social distancing detector application, which is an application that is used to detect the number and distance of human objects in one area. This study aims to develop a social distancing detector application using the Python programming language with the YOLOv3 library. YOLOv3 has advantages in object detection with high accuracy, which is above 90%. Testing the method in this study used five pedestrian datasets from road surveillance cameras obtained from trial datasets by five researchers through GitHub which have good resolution and have heterogeneous human objects. The result of the accuracy of the first image detection is 83.32%. The result of the accuracy of the second image detection is 76.92%. The result of the accuracy of the third image detection is 90.20%. The accuracy result of the fourth and fifth image detection is 100%. The average success rate of all analysis results is 90.08% as measured by the average comparison of the number of successful experimental data and the number of observational data for each image. © 2023 Author(s).
Department of Electrical Engineering, Faculty of Engineering, Universitas Negeri Semarang, Indonesia; Department of Electrical Engineering, Faculty of Engineering, Universitas Negeri Malang, Indonesia
