<![CDATA[Currently the counting survey is still being carried out manually using surveyors. The challenges include the need for high concentration, energy-draining tasks, and the requirement for a considerable number of surveyors, which are drawbacks in manual data collection. An approach can be taken by fully replacing it with the utilization of artificial intelligence. Using deep learning, research is conducted to design an automatic vehicle detection system by employing the YOLOv8 algorithm as a real-time based vehicle detection. Then, an analysis is performed to test the model's consistency in detecting six classified vehicle objects passing through one of the CCTV videos in Balikpapan's Transportation Agency (Dishub). Based on the analysis, the system's performance is obtained with the following accuracy rates: highest during the day, 96.92%–100%; lowest at night, 91.43%–100%. F1-Score values: highest at night, 80%–100% then in the morning and during the day, 67%–100%.]]>
Copyrights © 2025
