Pattern Matching based Vehicle Density Estimation Technique for Traffic Monitoring Systems
Sakthidasan Sankaran
Department of Electronics and Communication Engineering,
Hindustan Institute of Technology and Science, India
This email address is being protected from spambots. You need JavaScript enabled to view it.
Abstract: Due to increase in vehicle density, the road traffic estimation aids in enhancing the traffic management centre’s performance and their applications. The analysis of traffic surveillance based on video is an active research area that has varied range of applications in Intelligent Transport System (ITS). In specific, urban environments are much more challenging on comparing highways due to the placement of cameras, vehicle pose, background clutter, or variation orientations. There were several techniques employed so far for the process of traffic monitoring using pattern matching, however there were some limitations like reduced rate of accuracy and increased error rate. So as to overcome this, an efficient method is proposed. The main intention of this proposed approach is to monitor the density of traffic and to estimate the vehicle density using Pattern Matching for Vehicle Density Estimation (PMVDE) scheme. In this paper, the pattern matching based vehicle density estimation is employed for enhancing the detection of accuracy thereby reducing the rate of error. The region of interest of an image that is extracted from the video input is being analysed by this process. These two processes are employed in region of interest extracted image for decreasing the density detection errors. This approach attains less false positive rates and error rate, however this in turn influences the accuracy, precision, recall, F-score, and true positive rates and offers enhanced outcome on comparing other techniques.
Keywords: Pattern matching, traffic estimation, vehicle density, region of interest, feature extraction.
Received August 25, 2020; accepted June 16, 2021