An integrated system for tracking and recognition using Kalman filter

Abstract

Tracking of any object from a video scene becomes more critical not only for security applications but also for analyzing traffic. This system integrates low level image processing as well as high level image processing through online data model and offline data model for more efficiency and robustness. Also this makes the system to handle occlusion conditions and abrupt color intensity variation conditions. This system uses median filtering and blob extraction for moving object detection. The offline model employs high level image processing recognize the moving objects. Here the Kalman filter is used for effective tracking under complex situations.