Intelligent Video Surveillance Analytics with ADI’s Blackfin Processor

The market for video surveillance keeps burgeoning. Here’s a review of smart video surveillance technology and the challenges for embedded system designers, and an example of intelligent surveillance design using the Blackfin processor to provide the control and image processing.

By Harry Wei, Senior Technical Application Engineer – DSP/Embedded Processor, Analog Devices, Inc. and

Michael L. Long, Product Line Manager – Industrial Video and Imaging Solutions, Analog Devices, Inc

Page 1 of 5
Video/Imaging DesignWire
(10/16/2009 1:00:45 AM)

The modern video surveillance system has evolved from original analog CCTV (closed-circuit television), through the phases of digitization and networking, to a network of distributed video capture and analysis capable cameras. Advances in video compression technology have facilitated the digitization of the video surveillance system, reducing both large amounts of bandwidth needed for transport and overall storage capacity needed for archiving video.

The development of inexpensive wired and wireless computer networks with bandwidth to handle compressed video has brought surveillance on a campus and metro scale to reality.  The research into machine vision for industrial applications has developed recognition and tracking capabilities that turn a passive surveillance function into an intelligent security system that can recognize potential threats to safety and dangerous or suspicious activities.

Since the mid 1990s, real-time video analysis has made rapid advances, propelled by the research efforts of the momentous Visual Surveillance and Monitoring (VSAM) Project — sponsored by the U.S. Advanced Research Projects Agency (DARPA), with Carnegie Mellon University (CMU) and Massachusetts Institute of Technology (MIT) as the leading representatives, with participation by multiple American universities and other research facilities.  After the “9/11″ terrorist attacks on the U.S. and the explosives attacks on the Madrid train station and the London Metro, the worldwide demand for surveillance systems has seen unprecedented growth.

More than 4.2 million cameras have been installed all over the U.K., averaging one unit for every 14 persons, such that the average person would be observed by up to 300 units per day (The Daily Mail, U.K.). In China, the Guangzhou metropolitan area had installed 250,000  security cameras by the end of 2007, and the Beijing metropolitan area has installed public image information systems interconnected to the police surveillance networks in important facilities, densely populated public zones, traffic bottlenecks, vital civic infrastructure and the import zone, in addition to the 263,000 units already installed. Shanghai will install more than 200,000 surveillance cameras on the streets by 2010, and establish a “social preventive and controlling system”.  These massive surveillance efforts require video systems to intelligently compress, store and search the captured content.

Currently, besides the CMU and MIT effort, the Embedded Smart Camera Group from Graz University of Technology, Austria and the S3 Team of IBM, Irisnet (Internet-scale, Resource Intensive Sensor NETwork Services) and Intel have leading positions in the individual domains of distributed smart surveillance systems. Companies like Object Video, Hisign, and 3VR, are leaders in realizing the industrial application of smart video surveillance. In China, the Institute of Automation of Chinese Academy of Science, the Department of Electronics Engineering and the Department of Automation at Tsinghua University in Beijing are taking a leading position in research of this area.

NEXT: Smart Video Surveillance Technology

Page 1: next page

Pages: 1 2 3 4 5