Hard Sensor Processing for Data Fusion

Open Access
Agumamidi, Rachana Reddy
Graduate Program:
Electrical Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
May 11, 2011
Committee Members:
  • David J Hall, Thesis Advisor
  • Kenneth Jenkins, Thesis Advisor
  • Kultegin Aydin, Thesis Advisor
  • Data fusion
  • surveillance
  • video
  • tracking
There is a vast amount of information available these days that can be used in the right direction to prevent several extreme/disastrous events. This information is available in various forms such as images, videos, textual reports and other sensor outputs from deployed sensors such as video surveillance cameras, as well as from observations via mobile phones. All of these sensor outputs can be combined to help us perceive information about the environment around us. One of the important challenges in using all of these sensors is that each situation employs a different set of hard sensors and processing algorithms. There is no single algorithm and sensor architecture that can be used universally for all the different scenarios regarding suspicious activity recognition. Each of these scenarios is unique and employs a unique set of sensors and algorithms that must be applied separately. In this thesis, I have worked on some of the hard sensor processing algorithms and the architectures that can be utilized for multi-sensor data fusion applied to counterinsurgency situations and surveillance.