Real-Time Vehicle Detection from Captured Images

Objective of this paper is to primarily detect moving vehicles from real-time captured images so that we can later use them for a variety of purposes like detecting the number of vehicles that are present at a road at a certain point, understanding the motion of the vehicles, analyzing vehicle density at any traffic signal at any specified point of time; and based on our previous knowledge can make a good assumption where the vehicle will be heading towards. Signal-to-noise ratio, PSNER and mean-square-error are computed for each captured image using after detection. After establishing vehicle detection scheme, we can also make an attempt to read vehicle plate numbers further a much broader attempt to automate a traffic flow system using image recognition and probability distribution.