This study aims to propose a system that is able to detect crowds that have the potential to violate the Covid-19 health protocol. In this case the camera is used to capture objects in the form of images (image). Furthermore, the obtained images are processed using Deep Learning. In this case, the Convolutional Neural Network (CNN) is used. The criteria used are the classification of the number of people in the picture and the distance between each person in the picture. The distance allowed between each person is 1 meter. If there are more than 5 people in the image that have a distance between each of them less than 1 meter, then this is classified as a crowd that has the potential to violate the Covid-19 health protocol. In this study, 2 classifications were used, namely crowd and non-crowded. From the test results obtained the average value of accuracy, precision, and recall are 91.71, 91.25, and 92.65, respectively.
Copyrights © 2022