<![CDATA[Traffic density on the highway, especially at intersections where there are traffic lights, is a number that can be used to estimate the number of vehicles. The increase in the density of highways is directly proportional to the growth in Indonesia's population which is increasing. The problem was found that the survey to record the types of vehicles in Indonesia was still manual by taking notes and coming directly to the location. Based on this problem, a vehicle object detection system was created, especially cars based on their brands and variants to facilitate the police survey process to detect violations that have occurred. This study uses the You Only Look Once (YOLOv4) algorithm to detect cars, and classify and determine the level of accuracy. This research uses 7034 image dataset with 8 classes, namely Motorcycle, Suzuki Ertiga, Honda Jazz, Honda Brio, Toyota Avanza, Toyota Innova, Mitsubishi Xpander, Mitsubishi Pajero Sport. The results of this study indicate that the YOLOv4 algorithm can be used on Jogja CCTV to detect vehicle types with an accuracy of 82%.]]>
Copyrights © 2024
