Cyclops, image sensing and interpretation in wireless networks

  • Authors:

  • Mohammad Rahimi;Deborah Estrin;Rick Baer;Henry Uyeno;Jay Warrior

  • Affiliations:

  • University of California at Los Angeles, Los Angeles, CA;University of California at Los Angeles, Los Angeles, CA;Agilent Laboratories, Palo Alto, CA;Agilent Laboratories, Palo Alto, CA;Agilent Laboratories, Palo Alto, CA

  • Venue:
  • SenSys '04 Proceedings of the 2nd international conference on Embedded networked sensor systems
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Technological progress in integrated, low power CMOS imaging [1] and maturing low power wireless sensor network platforms, motivate a new and rich design space exploiting in network dense imaging. In particular, by combining comparably low power CMOS cameras and low power wireless sensor nodes, and implementing on board compression and image analysis techniques, we can greatly enhance the application of dense wireless sensor networks to phenomena that are most readily observed in the optical domain. In this demonstration we will introduce Cyclops. Cyclops (see Figure 1) is a joint project between Agilent Technology and the Center for Embedded Networked Sensing (CENS). It is a low power CMOS imager with a local frame grabber and local computation. It behaves like a sensor for a host and can do local computation and inference, acting as a more capable sensor. Inference can be based on a color histogram, statistical characteristics of images, or a hypothesis such as the existence of motion or a specific template in the scene. The demonstration will showcase the functionality of the first generation of Cyclops, including low duty cycle image capture and participation in a sensor network via an attached Crossbow Mica2. Performance studies will also be presented, along with outstanding challenges and foreseeable opportunities.