Real-Time Tracking of Multiple Moving Objects in Video Using Proximity

Abstract

Daily experience tells us that human visual systems
are very fast and intuitively recognize and track objects around
us. Real-time detection and tracking of a changeable number of
objects is of high interest for several applications such as visual
surveillance, traffic monitoring, human computer interaction etc.
This paper presents a novel method for real time object detection
using background learning and tracking by proximity analysis of
the moving objects. We present our system for multiple moving
object detection and tracking using webcam installed inside a
building that monitors a typical open work area like lab, hospital
or parking area. Object detection is the first and most important
step of moving object tracking. Two main processes are implemented
in this system. The first process used a simple technique
to extract initial background model from first N frame taken
from the camera sensors. This helps in detecting moving objects
from background. The second process continuously updates the
incoming frame for object depending upon the proximity analyses
to check whether he object is old, new or removed from the frame
sequence with respect to the background. Videos from webcam in
indoor and outdoor environments are used to generate the result.
We have implemented a new method of multiple objects tracking
by using proximity rules of grouping for handling the general
problem of multiple moving objects, in various environmental
conditions.