Sistem Tracking dan Counting Kendaraan Berbasis YOLO untuk Pemetaan Slot Parkir Kendaraan

Abstract

The growth in the number of vehicles in Indonesia has always increased from year to year. Based on data from the Central Statistics Agency, it shows that from 2018 to 2020 there has been an increase of around 5.62% (2018-2019) and 1.89% (2019-2020). This condition causes the condition of parking lots in every public facility to always be a concern considering the lack of availability of parking spaces accompanied by an increase in the number of vehicles. This causes, in some places, especially for areas such as malls, the parking area is always filled with vehicles. Finally, it caused several visitors to experience difficulties in parking the vehicle. In this study, developing an automatic parking system by detecting, tracking and counting vehicles in real time. The proposed detection system is based on YOLO and to carry out vehicle calculations using calculations by checking vehicles that have crossed a predetermined center line. The system performance results show that the detection results show performance with an accuracy percentage of 85% and the results of counting or object calculations with an accuracy of 77.55%.