Centroid Tracking and Target Identity Estimation using Image Sensor Data

Naidu, VPS and Girija, G and Raol, JR Centroid Tracking and Target Identity Estimation using Image Sensor Data. Technical Report. NAL.

[img] PDF
PD0503.pdf

Download (322kB)

Abstract

Algorithms for centroid tracking and target identity estimation using image sensor data are implemented in PC MATLAB. For tracking using image sensor data, the centroid of images is determined by using segmentation technique. This is achieved by converting the gray level image into binary image and reduced to clusters by nearest neighbor (NN) criterion. The performance of the algorithm for centroid tracking has been evaluated using simulated data. Bayesian sensor fusion technique is used for estimation of target identity and location using simulated data of an imaging sensor and an acoustic sensor. This is achieved by computing the likelihood functions for the two sensor data and fusing the posterior probabilities. From the posterior probability distribution, the target identity and location are established using search algorithms. A comparison of three search algorithms: Metropolis Hastings, Simulated Annealing and Gradual Greedy is made for a two dimensional scenario with two targets.

Item Type: Monograph (Technical Report)
Subjects: AERONAUTICS > Aircraft Communication & Navigation
AERONAUTICS > Avionics & Aircraft Instrumentation
ENGINEERING > Communications and Radar
Depositing User: Dr VPS Naidu
Date Deposited: 26 Jul 2010 08:07
Last Modified: 28 Mar 2011 07:41
URI: http://nal-ir.nal.res.in/id/eprint/8532

Actions (login required)

View Item View Item