Real time web vehicle classifier

This paper discusses the issues in the design and implementation of a Real Time Web Vehicle Classifier system that is an automated intelligent pattern matching system. It makes vehicle classifier information and video images accessible over the Internet in real time. The classifier system uses image-processing techniques to classify a vehicle's type from video frames. The classifier system processes video images frame by frame and displays the classified information on host machine's screen. At the same time both the video and the classification data are stored into a real-time database for on-line searches and streaming video players. Potentially, the Real Time Web Vehicle Claffifier is convertible into other types of surveillance or survey applications over the WWW. This paper explains the system architecrue and shows how Canny style derivative edge detection is used to classify vehicles and how image files are stored into database, and how the vehicle classifier information is displayed on client machines in real time