University of California, Riverside UCR


CRI: Outdoor Video Sensor Network Laboratory

Supported by National Science Foundation grant 0551741.

Principal Investigator

Bir Bhanu, Center for Research in Intelligent Systems EBU2 Room 216, University of California at Riverside,
Riverside, CA 92521, Tel. 951-827-3954, Fax. 951-827-2425
bhanu@cris.ucr.edu
http://www.vislab.ucr.edu/PEOPLE/BIR_BHANU/index.htm

Co-PIs

Amit K. Roy-Chowdhury, Center for Research in Intelligent Systems , Dept. of Electrical Engineering,
University of California at Riverside, Tel. 951-827-7886, Fax. 951-827-2425
amitrc@ee.ucr.edu
http://www.ee.ucr.edu/~amitrc/

Chinya Ravishankar, Center for Research in Intelligent Systems , Dept. of Computer Science and Engineering,
University of California at Riverside,Tel. 951-827-2451, Fax. 951-827-4643
ravi@cs.ucr.edu
http://www.cs.ucr.edu/~ravi

Students

Ramiro Diaz, Ankit Patel, Hoang Nguyen, Mostafa Elhams, Huy Tran, Mauro Ibarra


Research and Education Activities

Publications and Products

  • J. Yu, B. Bhanu, Y. Xu and A. Roy Chowdhury, "Incremental Construction of Super-resolved 3D Facial Texture in Video," International Conference on Image Processing, 2007.
  • X. Zou, B. Bhanu, B. Song and A. Roy Chowdhury, "Determining Topology and Identifying Anomalous Patterns in a Distributed Camera Network", International Conference on Image Processing, 2007.
  • B. Song, A. Roy-Chowdhury, “Stochastic Adaptive Tracking In A Camera Network,” IEEE Intl. Conf. on Computer Vision, 2007.
  • B. Song, N. Vaswani, A. Roy-Chowdhury, "Closed-loop Tracking and Change Detection in Multi-Activity Sequences", IEEE Conference on Computer Vision and Pattern Recognition, 2007
  • X. Zou and B. Bhanu, "Anomalous activity classification in the distributed camera network," International Conference on Image Processing, San Diego, CA, Oct. 12-15, 2008.
  • J. Yu and B. Bhanu, "Super-resolution of facial images in video with expression changes," 5th IEEE International Conference on Advanced Video and Signal Based Surveillance, Sept. 1-3, Santa Fe, New Mexico., 2008.
  • J. Yu and B. Bhanu, "Super-resolution of deformed facial images in video," IEEE International Conference on Image Processing, San Diego, CA, Oct. 12-15, 2008.
  • Y. Li and B. Bhanu, “Utility-based dynamic camera assignment and hand-off in a video network,” Second ACM/IEEE International Conference on Distributed Smart Cameras, pop. 1-9, Stanford, CA, Sept. 7-11, 2008.
  • B. Song and A. Roy-Chowdhury, “Robust Tracking in A Camera Network: A Multi-Objective Optimization Framework,” IEEE Journal on Selected Topics in Signal Processing: Special Issue on Distributed Processing in Vision Networks, August 2008.
  • B. Song, C. Soto, A. Roy-Chowdhury, J. Farrell, “Decentralized Camera Network Control Using Game Theory,” Workshop on Smart Camera and Visual Sensor Networks at IEEE/ACM Intl. Conf. on Distributed Smart Cameras, 2008.

This NSF project develops a new laboratory and conducts research in video understanding and related technologies in a wireless network environment. While research into large-scale sensor networks is being carried out for various applications, the idea of massive video sensor networks consisting of stationary and moving cameras connected over a wireless network has been largely unexplored. Wireless video sensor networks are necessary for a number of life-critical applications such as surveillance for homeland security, scene analysis of disaster zones for coordinating rescue efforts, wildlife monitoring and the entertainment industry. Wireless sensor networks have the crucial advantage of mobility and ease of installation of sensors, but suffer from power and bandwidth constraints. Video processing and transmission require large amounts of computing power and transmission bandwidth. So one is concerned with these trade-offs. The proposed laboratory, under development, will provide a state-of-the-art facility for conducting research and teaching. It consists of 80 pan-tilt-zoom video cameras that can be accessed over the network using an IP address. Each camera is connected to a computational unit that takes care of local processing at the sensor node. It identifies the data in the video sequence relevant for a particular application, which is then compressed and transmitted. This object-based distributed compression scheme significantly reduces the bandwidth requirement from the network. In order to save battery power at the sensors, a triggering mechanism based on acoustic, seismic and vibration sensors is used. A few infrared sensors will supplement the data provided by the color video cameras to perform diurnal scene analysis. Some of the sensors are fixed and powered by connecting to an electrical outlet, while the mobile ones are powered from solar energy.