A feature-based vehicle tracking system in congested traffic video sequences

Abstract

This paper describes a new feature-based vehicle tracking system using trajectory matching, which extracts corner features of the vehicle and tracks the features using linear Kalman filtering, where features from the same vehicle are grouped together. We also propose a new grouping algorithm using trajectory matching to make our tracking system robust enough for segmenting different vehicles in the congested traffic situation. The proposed system has demonstrated good performance for crossway traffic video sequences.