Is There a Market Opportunity for MV Companies?
Introduction
Close circuit video surveillance systems are
nearly ubiquitous today in vulnerable locations. First introduced in the
1960æ? video surveillance systems have become central tools for providing
security for critical infrastructure, institutions and establishments; including
cargo ports, dams, airports, trains and subways, financial and governmental
institutions, communication networks, power grids and oil transport and refining
facilities. Video surveillance systems have also become widespread among
businesses with entranceways, walkways and backdoors closely monitored by video
cameras. In banks, surveillance cameras are strategically positioned to monitor
and/or record the behavior of customers and tellers. Even corner convenience
stores have embraced the use of video surveillance systems.
Conventional Versus High-End Surveillance Systems
In many
cases, the systems deployed involve a relatively simple arrangement. A
stationary camera continuously scans a scene within its field of view. The
signal from the camera is then transmitted and recorded on a VCR, or observed in
real-time on a monitor by security personnel, or both.
This conventional arrangement is highly limited in its capabilities. VCR
tapes must be changed frequently and typically are saved for limited periods and
then erased through re-recording. In larger scale deployments, security
personnel sit before a bank of video monitors, attempting to detect threatening
behavior. However, as human ability to process multiple video signals over
extended periods is limited, the senses are easily overwhelmed with resultant
attention lapses entailing potentially serious security breaches. Additionally,
conventional video surveillance systems can suffer from technical deficiencies
as in the case of low light conditions.
Fortunately, the state of the art of video surveillance is changing. In
particular, the horrific events of 9-11 have underlined the need for more
capable, higher-end video surveillance systems, providing an impetus for the
development and deployment of such systems. At the same time, technological
advancements have made it possible to answer this need. These advancements
include:
- Thermal imaging devices for low light conditions (primarily IR and Near-IR
cameras)
- Remotely controlled PTZ (point, tilt and zoom) cameras
- Motion detection embedded in cameras to limit video signals to events worthy
of observation and recording (Cameras begin scanning when triggered by motion,
which limits the bandwidth needed for transporting video signals on the security
network.)
- Digital cameras in place of analog cameras (Digital video signals can more
readily be transported over IP networks, processed by software and recorded on
digital video recorders. Most importantly, digital video data can be processed
by computers in contrast to analog video signals.)
- IP addressable cameras to take advantage of Internet connectivity (IP
transport of video signals offers far greater flexibility than traditional,
point-to-point, analog networks. Once linked to an IP network, cameras can
become part of a much more broadly deployed system where video feeds from
distant locations can be recorded and analyzed centrally.)
- Digital video recorders (DVRs) which record video signals on hard drives
instead of VCR tapes
- Sophisticated image processing software for system control, content
analysis, retrieval and decision-making
Higher-end video surveillance systems (a.k.a. intelligent Video Surveillance Systems? for short) make use of these technological innovations. The
central defining difference between conventional and higher-end video
surveillance systems, however, lies in the way video signals are processed. In
the case of conventional systems, humans process the signals with their eyes and
brains and then determine appropriate responses. By contrast, IVS utilizes
sophisticated software to process video signals for decision making. IVS is thus
largely an automated, software-driven approach to surveillance.
How IVS Works
As with any surveillance system, the
starting point with IVS is specifying the entity to be secured, such as an oil
refinery, and the parameters of the physical area to be monitored, such as a
perimeter fence. Threatening scenarios are then predetermined by level of
severity and programmed into the system. Examples of such scenarios could
include individuals or vehicles approaching or running near the fence,
individuals appearing to be observing the facility from outside the fence
(using the joint? and objects abandoned near the fence (such as high explosives
in a suitcase). In effect, the surveillance system ? to look for
specific anomalies, that is; activities or objects that depart from a normal,
non-threatening baseline. When an anomaly is detected, the system can respond
by automatically directing a PTZ camera to zoom in or investigate further from a
different angle or camera. Where a threshold level is met, an alarm is triggered
and transmitted to security personnel, enabling a real-time, protective
response.
Importantly, IVS does not require the existence of a physical perimeter. In
fact, the sophisticated software that comprises the core of an intelligent
surveillance system can create virtual perimeters across land and water by means
of a video tripwire at is drawn on a computerized camera snapshot.
To realize this capability, the system must first extract a exta-data? stream
from real-time CCTV video that captures the activities of every object in the
video stream. This is not video in the conventional sense but instead
information about the raw video feeds that have been cleansed of redundant and
irrelevant content. At the same time, the raw video feeds are compressed and
archived for future retrieval.
To detect anomalies, IVS runs metadata through content analysis and inference
engines based on sophisticated motion detection and image recognition
algorithms, involving blob extraction by background modeling or frame
differencing, blob or feature tracking from frame to frame; object
classification by shape, size, color; and algorithms for determining the
activities of individuals.
Beyond real-time applications, the use of metadata in IVS enables
analysis? where images stored on DVRs can be easily and quickly searched to
examine activities occurring before and after alarms are sounded in an attempt
to reveal the identities of perpetrators. Utilizing a user-friendly GUI, a
security official, for example, could issue a command to the system, such as: for a red SUV lingering before an airport terminal entrance.?nbsp; Or,
because surveillance information can be stored centrally from multiple, remote
locations, the same security official could perform a search for the same SUV at
different airports in an effort to detect a pattern, signifying a potentially
more widespread threat.
The Future of IVS and Market Opportunities for Machine Vision
Companies
The future bodes well for IVS. Threats to strategic
assets and to locations where large groups of people congregate are not likely
to diminish soon, while -at the same time- the inadequacies of conventional
security systems are painfully clear. A proliferation of intelligent,
software-driven security systems is therefore likely. Based on the fusion of
camera and information technologies, these systems will greatly magnify the
ability of security personnel to protect society against serious threats.
Not surprisingly, large though varying- market opportunities are foreseen
for IVS:
- J.P. Freeman has forecast the size of the overall video surveillance market
at $3 billion (USD) by 2007.
- Axis Communications estimates the US security market at $4 billion (USD).
- In-Stat MDR has placed the number of surveillance cameras sold in 2004 at
15.2 million, while ObjectVideo estimates annual sales of security cameras at 6
million.
Does this market opportunity lend itself to machine vision companies? The
strong reliance of IVS on image capture and processing technologies suggests a
good fit for machine vision companies. In particular, manufacturers of cameras,
optics, imaging boards and software appear well positioned to reap benefits from
IVS. As this illustrates, and as noted in previous articles, market
opportunities for machine vision companies appear to lie increasingly beyond the
factory floor. The good fortunes of the machine vision industry can only
expand, as more and more companies seek and realize these new opportunities.
HIGH-end Video Server Module board W1860
