Vehicle capturing and counting using a new edge extraction approach

This paper presents a new edge extraction approach to detect a moving vehicle in a monocular video sequence. The new edge extraction improves the tolerance of the detection process toward variations in the illumination conditions and decreases the influence of the vehicle cast shadows. In terms of their edge information, the vehicle detection based on calculating the variance of edge information and nighttime algorithm will be introduced. Our method specifies the rectangle area of moving vehicles based on one-lane road which are able to support the real-time processing of image sequences on DSP.