AUTOMATED QUEUE ANALYSIS FOR SERVICE MONITORING USING COMPUTER VISION

• Main Author: Pradipta Andila W., Rheznandya
• Format: Final Project
• Language: Indonesia
• Subjects: Teknik (Rekayasa, enjinering dan kegiatan berkaitan)
• Online Access: https://digilib.itb.ac.id/gdl/view/82300
• Institution: Institut Teknologi Bandung

In public services such as airports, the length of the queue is used as a performance indicator since it impacts customer satisfaction. However, queue monitoring is often still performed manually thus limiting the extent of the monitoring process. This study aims to develop an automated system that monitors queues that are capable of continuous analysis. The method employed in this study involves creating a prototype of a queue analysis system using computer vision. The technology utilised is YOLO (You Only Look Once) for object detection and a Python program to calculate key parameters such as service time, service rate, interarrival time, arrival rate, and utilisation factor. The prototype results show that the automatic calculations, compared to manual calculations, have an average service time error percentage of 20.41%, and an average error for the arrival rate of 41.72%. This discrepancy is due to the system still using YOLOv4 as is, without any retraining. Although the error rate is still high, the system generally demonstrates correct characteristics in depicting queue conditions and can be improved by retraining YOLO weights specifically for human detection. This study lays the foundation for the development of an automatic queue analysis system beneficial for maintaining public service levels at airports.