Vehicle Activity Analysis Based on ANPR System

Abstract

Automatic Number-Plate Recognition (ANPR) system is commonly deployed on the road networks and used in field of safety and security systems. It is interesting to figure out the purpose of each vehicles based on the huge amount of ANPR data captured daily by embedded ANPR cameras. In this paper, we propose spatial and temporal quantitative indicators of vehicle trace features and the extraction algorism. We applied data-mining techniques to analysis vehicle activity patterns from ANPR data, and proposed centroids based activity classification method. Evaluations show the capability and efficiency of the proposed approach.