Abstract

The increase in vehicular traffic demands an effective and well-organized solution to manage and regulate traffic. For recent advances in Intelligent Transportation Systems (ITS), the video-based traffic monitoring system is the most popular and practical solution. The major component of such systems is the vehicle detection process. The accuracy of the existing detection approaches depends on the effectiveness of the image processing technique used, but it should be noted that it depends primarily on section of the scene in the video being processed. This prerequisite task of selecting the significant area of video is ignored which can elevate the accuracy of detection to a different level. This paper proposes a detection framework specific to vehicle detection systems known as Virtual Mono-Layered Continuous Containers (VMC2), which simplifies the entire detection process. The performance of the proposed framework is analyzed using existing popular methods of detection based on background subtraction such as running average; Gaussian mixture (MOG) and K-Nearest Neighbor (KNN). The results in terms of computational complexity and storage space are presented. From the results it is seen that the proposed framework significantly reduces the false positives leading to a highly accurate vehicle detection system.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call