Deteksi Orang Tenggelam Di Kolam Renang Umum Secara Realtime Menggunakan Metode Ssd Mobilenet V2

Abstract

Swimming is now an activity or sport that is in great demand by many people ranging from children to adults.Swimming is an active and healthy activity, but few people pay attention to safety when participating in these activities.Although most swimming pools have professional guards, there are still many areas that are not covered by supervision and result in drowning accidents while swimming due to late rescue.drowning accidents in swimming pools in This study focuses on the discussion of safety systems in swimming pools that can provide fast information to lifeguards so that they can rescue in the event of an accident. A digital image processing system is needed to support the detection process in this research, a system that has fast and accurate performance is needed. The author uses the SSD-Mobilenet V2 method to detect movements or indications of drowning accidents in swimming pools in realtime. Detection is carried out on movements that indicate a person or pool visitor who is about to drown, and uses CCTV cameras at a distance of 2-4 meters and a depth of 2- 5 meters. The test results show that the system using the SSD-MobileNet V2 method is able to detect movements that indicate drowning with an accuracy of 87%.