Abstract
In the field of traffic engineering, video image analysis is an essential technique for correct understanding of traffic phenomena and has been often used for many purposes. However, manual use of this type of technique is a very labor-intensive work. In this study, a multi-functional video image processing system is developed, which can 1) count traffic volumes, 2) track moving objects such as vehicles and pedestrians, and 3) recognize license plates automatically by using image processing technology. Case studies are carried out by applying this system to the actual road traffic, and its performance and accuracy are evaluated. In conclusion, the developed system showed sufficiently satisfactory performance for practical use.
