An Effective Calculating Method of Signalized Intersection Delay

Delay of signalized intersection has been estimated not only by manual method but also some intersection models. Long queues of lanes will make it difficult for collecting data manually and delay calculated by signalized intersection models may not be perfect for a particular intersection. In this paper, we design a system that can calculate delay automatically based on image processing technology. Our system can detect vehicles by placing two detection areas that one is upstream further than the last vehicle waiting in the queue and the other near the stop line and then track each vehicle not only in one lane but also in several lanes. Then it just calculates each vehicle delay and gets signalized intersection delay by averaging these data. The result indicates that it is effective and can meet real-time application.