Some Applications of Distributed Architectures in Image-based Surveillance Systems

1
Some Applications of Distributed Architectures in
Image-based Surveillance Systems

• Graduate Seminar in CIS 750 Video Processing and
  Mining Spring 2003
• Presented by
• Benjamin Garrett

2
Agenda

• Distributed Multi-Sensor Surveillance Issues and
  Recent Advances
• Pramod K. Varshney and Ioana L. Coman
• Sensing Devices
• Architectural Issues
• Information Processing
• Case study concealed weapons detection

3
Distributed Multi-Sensor Surveillance

• Primary aspects to take into consideration
• Sensing device development some added processing
  capabilities needed.
• System and network design global considerations
  and architectural issues need attention.
• Information processing tasks the abundance of
  data requires elaborate information fusion
  techniques.

4
Distributed Multi-Sensor Surveillance

• Intelligent Distributed Systems (IDS)
• Includes some work on specialized OS
• Intelligence Surveillance and Reconnaissance
  (ISR)
• Data processing issues arise due to large
  quantities of data
• Intelligent Real-time Integrated Sensor (IRIS)
  systems
• Redundant sensors for added reliability

5
Sensing Devices

• Early efforts were dedicated to the development
  of various types of sensing devices
  acoustic/sonar, IR, seismic, magnetic.
• One example is the Remote Battlefield Acoustic
  Sensor System (REMBASS)

6
REMBASS

• Ground-based, all-weather, day-and-night,
  battlefield surveillance, target development, and
  early warning system capable of remote operation
  under field conditions.
• Basic purpose of REMBASS is to detect, locate,
  classify, and report personnel and vehicular
  (wheeled and tracked) activities in real-time
  within the area of deployment.
• It uses remotely monitored sensors placed along
  likely enemy avenues of approach.

7
REMBASS

• Sensors respond to seismic-acoustic energy, IR
  energy, and magnetic field changes to detect
  enemy activities.
• The sensors process the data and provide
  detection of classification information which is
  incorporated into digital messages and
  transmitted through short burst transmission to
  the system sensor monitor programmer set.
• The messages are demodulated, decoded, displayed,
  and recorded to provide a time-phased record of
  enemy activity.

8
REMBASS - Problems

• Sensors had limited processing power and
  over-loaded the central unit with data.
• Limited bandwidth and information fusion
  capabilities at the central unit did not allow
  optimum utilization of the retrieved data.

9
MEMS

• Micro-Electro-Mechanical Systems (MEMS)
• integration of mechanical elements, sensors,
  actuators, and electronics on a common silicon
  substrate through micro-fabrication technology.

10
MEMS

• Promises systems-on-a-chip capabilities.
• Microelectronic integrated circuits can be
  thought of as the "brains" of a system and MEMS
  augments this decision-making capability with
  "eyes" and "arms", to allow microsystems to sense
  and control the environment.
• Sensors gather information from the environment
  through measuring mechanical, thermal,
  biological, chemical, optical, and magnetic
  phenomena.
• The electronics then process the information
  derived from the sensors activate mechanical
  devices.

11
The Georgia Tech Wearable Motherboard

• Promises a multitude of applications in sports
  medicine, advanced health care, and monitoring of
  astronauts, law enforcement personnel, and combat
  soldiers.
• Optical fibers can detect bullet holes, and
  special sensors and interconnects monitor vital
  signs of the body.

12
The Georgia Tech Wearable Motherboard

• Plastic optical fibers woven throughout the
  fabric of the shirt.
• Flexible data bus transmitting information from
  sensors mounted on an inside shirt.
• Bus also transmits information to the sensors
  (and hence, the wearer) from external sources.
• The optical fiber can be used to pinpoint the
  location of a bullet penetration in combat
  causality care.

13
Distributed Multi-Sensor System Architecture

• Operational Independence
• Managerial Independence
• Evolutionary Independence
• Emergent Behavior
• Geographic Distribution

14
Distributed Multi-Sensor System Architecture

• Sensor-level intelligent subsystems one or a
  few devices configured for fast reaction time.
• Regional or local subsystems where data fusion
  takes place.
• Central Intelligence Units usually few if not
  only one. Makes complex decisions and can
  override decisions of lower level units.

15
Distributed Multi-Sensor System Architectural
Issues - IDS

• Encompasses a wide range of activities.
• Intelligent Interactive Distributed Systems group
  - Vrije Universiteit (VU) in Amsterdam
• Agent Operating System a platform for managing
  mobile processes.

16
Information Processing

• Refers to effective means for coordinating the
  data coming from multiple sensors.
• Data/image/information fusion is a vast research
  field with many open projects in progress.
• Video Surveillance and Monitoring Team at CMU

17
VSAM at Carnegie Mellon
18
VSAM at Carnegie Mellon

• Data fusion Every observed object is positioned
  in a 3D geodetic coordinate system using
  geolocation.
• Sensor Tracking Sensors considered as precious
  resource to be allocated according to
  user-specified tasks.
• Scene Visualization Employs a GUI giving a
  synthetic view of the environment.

19
VSAM at Carnegie Mellon
20
Case study concealed weapons

• Uses two different types of sensors MMW and IR
  wave sensors.
• Infrared waves give better resolution.
• Millimeter waves penetrate better.

21
IR waves and Millimeter waves
22
Case study concealed weapons

• Image Registration The process of finding the
  corresponding points from two or more images.
• IR image is superimposed over the MMW to evaluate
  the accuracy of registration task.

23
Distributed Surveillance Systems Concluding
remarks

• High amounts of funding being invested in
  distributed multi-sensor surveillance systems.
• Many of the issues presented are open research
  problems, some of which are still in their
  initial stages of development.
• Encompasses a wide variety of disciplines and
  fields.

24
Sources Consulted

• 1 Bult K. et. al. Low Power Systems for
  Wireless Microsensors, Proc. of the 1996 Intl.
  Symposium on Low Power Electronics and Design,
  Monterey, CA, Aug. 1996, pp. 17-22
• 2 Lin T.-H., Sanchez H., Kaiser W. J. and Marcy
  H. O. Wireless Integrated Network Sensors (WINS)
  for Tactical Information Systems, Proc. of the
  1998 Government Microcircuit Applications
  Conference.
• 3 Sungmee Park, Kenneth Mackenzie, Sundaresan
  Jayaraman. The wearable motherboard a framework
  for personalized mobile information processing
  (PMIP). 170-174 Electronic Edition (DOI
  10.1145/513918.513961)
• 4 Babak Firoozbakhsh, Nikil Jayant, Sungmee
  Park, and Sundaresan Jayaraman. Wireless
  Communication of Vital Signs Using the Georgia
  Tech Wearable Motherboard, IEEE Intl. Conference
  on Multimedia Expo. 2000, New York, NY,
  Electronic Proceedings.

25
Sources Consulted cont.

• 5 Y. Wang and B. Lohmann. Multisensor image
  fusion concept, method and applications.
  Technical report, University of Bremen, 2000.
• 6 H. Qi, X. Wang, S. S. Iyengar, and K.
  Chakrabarty, Multisensor data fusion in
  distributed sensor networks using mobile agents,
  Proc. Intl. Conf. Information Fusion, pp. 11-16,
  August 2001.
• 7 R. Collins, A. Lipton, H. Fujiyoshi, and T.
  Kanade. Algorithms for cooperative multisensor
  surveillance. Proceedings of the IEEE, Vol. 89,
  No. 10, October, 2001, pp. 1456 1477.
• 8 The Intelligent Interactive Distributed
  Systems group web site http//www.iids.org/.
• 9 The Remote Battlefield Acoustic Sensor System
  web site http//www.fas.org/man/dod-101/sys/land/
  rembass.htm
• 10 The Micro-Electro-Mechanical Systems
  Clearinghouse web site http//www.memsnet.org/mem
  s/