analytics article aug09

8/6/2019 Analytics Article Aug09

http://slidepdf.com/reader/full/analytics-article-aug09 1/428 urgent communications august 09

Keen vision

The surveillance industry has changed dramati-

cally in recent years. Not that long ago, analog

camera systems with analog video-tape record-

ing systems were the industry standard. Then, digital

video recorder–based solutions were introduced, allow-

ing the connection o a limited number o analog and/

or digital camera eeds, using a proprietary hardware

platorm. Now, surveillance solutions are managed

through so-called network video recorders that eature

video management sotware applications operating on

standards-based computer hardware.

IP cameras are connected to the NVR server through

IP-based networks. Legacy (analog) cameras can betied into these digital surveillance systems using

encoders. The number o cameras

within a single system has virtu-

ally no boundaries, and complex,

distributed architectures can be

supported with multiple server

and storage locations. As long as a

high-speed connection is available,

any networked computer then has

access through the NVR servers to

any o the cameras or video moni-

toring and system management

purposes.

In parallel, broadband wireless

technologies have matured, while

the allocations o the 4.9 GHz band

or public-saety applications ueled

the adoption o citywide surveil-

lance systems to address saety

and crime eectively in medium to

large cities throughout the United

States. Since the wireless plat-

orms today are natively IP based,

these implementations logically aredigital and NVR-based implemen-

technology digital video

ay hy b-fy

wh w wy f

B l Kwjzk J Bz

Viewing/ OperatorStation

MonitorRoom

RemoteViewing

StorageDVMSServer

Server Room

Wireless Cameras

Wired Camera

AnalyticsServer

Viewing/ OperatorStation

AnalyticsServer

Internet/Intranet

Edge Implementation+

Central Implementation

Figure 1 edge implementation vs. central implementation


http://slidepdf.com/reader/full/analytics-article-aug09 2/4

29urgentcomm.com

tations. Typically, cities also look

toward integrating existing, legacy

systems into the NVR-based solu-

tion, such as in-building security

cameras systems.

As cities start building their uni-

ied, digital surveillance systems

and add more and more cameras,

one might wonder how this grow-

ing number o cameras can be

managed. Ater the change romanalog systems to open standards-

based scalable, digital solutions,

the next paradigm shit within the

surveillance industry very well

might be the shit to intelligent

surveillance, rather than current

evolutionary progress, such as the

emergence o H.264 encoding and

megapixel cameras.

Research has indicated repeat-edly that the concentration span

o surveillance system monitoring

personnel is limited. As the num-

ber o video eeds grows, the ability

o monitoring personnel to eec-

tively identiy suspicious behavior

declines. The purpose o video ana-

lytics now is to make surveillance

systems “intelligent,” but only to

a certain extent. There should not

be an expectation that the humanoperator can be replaced, nor that

all alarms generated rom the video

analytics engine can be considered 100% accurate. But

the idea is that such systems analyze the video eeds

and alert the operator o situations that would require

a closer look; a human assessment. The operator now

concentrates primarily on only those ew camera eeds

it is alerted to by the video analytics engine, and the

overload o the operator is no longer an issue.

The video analytics process is highly complex and

processing intensive. This is why basic video man-

agement solutions have not integrated any analytics

eatures beyond basic motion-detection. Specialized

video analytics packages are available as overlay

implementations to augment the basic video manage-

ment solutions.

Video analytics engines ollow similar steps in ana-

lyzing video. First, in the process o segmentation

— based on image changes rom rame to rame — the

video analytics engine identiies the oreground and

background pixels. The result o the segmentation pro-

cess is the identication o a certain number o “blobs,”

with a blob being a collection o connected pixels.

Next, the process o classication assigns a class to

each identiied blob. Given a variety o applications,classes might include people, certain types o vehicles,

animals and certain static objects.

As classiied blobs move through the ield o view,

in multiple rames, tracking algorithms are used to

ollow each blob. By analyzing the class o the blobs

in combination with the tracked movement pattern,

activity recognition can be derived to identiy the

behavior o the blobs, and more importantly, possible

suspicious behavior. I multiple people blobs converge,

we can conclude that a crowd is orming. I a vehicle

blob stops moving in an area where parking is prohib-ited, this vehicle is illegally parked. Typical behaviors

that can be identied with analytics engines include

crowd orming, object removed/let behind, crossing o

perimeter lines and loitering.

Several years ago, due to processor intensiveness,

a central video analytics server would be able to

handle only very ew camera eeds in terms o ana-

lytics, making solutions expensive and impractical

to implement. As a result, the concept o distributed

analytics processing at the “edge” (i.e., with the cam-era unit) became popular to allow or better scalability.

However, edge-based analytics processing introduces

Acquire

Analyze

Alert

Assess

Act

s: c

5 steps For BetterVideo surVeillance



30 urgent communications august 09

the issue o compatibility. For new

deployments, only certain encod-

ers or IP cameras that allow or

running a certain analytics solu-

tion can be used. Furthermore, or

existing camera systems, signii-

cant upgrades might be required to

enable the edge processing.

Based on a review o current

video analytics solutions, given the

strong advances in hardware pro-

cessing power, the central processor

approach today supports many

more cameras per server, compared

just a ew years ago. But strong vari-

ance between solutions exists, with

a single server able to support any-

where rom 12 behaviors in total to

as many as 100 cameras with any

number o behaviors.

In terms o camera design, it

should be clear that current video

analytics technologies require still

images, so that the systems can

learn the static background (i.e.,

background pixels) and identiy

the oreground blobs. Regarding

ixed cameras, the image is inher-

ently “still,” as the camera will not

move, and the analytics add-on is

thereore straightorward. How-

ever, pan-tilt-zoom (PTZ) cameras,

commonly used in citywide surveil-

lance applications, might be used

in a tour mode o operation and

require urther consideration. With-

out adding additional cameras at a

location o interest, the tour mode

needs to be disabled when analytics

are applied to a speciic PTZ cam-

era. In other words, a PTZ camera

could be programmed to operate in

a static, preset position with analyt-

ics enabled during certain portions

o the day, while during other parts

o the day the PTZ camera would

resume its tour mode with analyt-

ics disabled. For instance, during

the day, a regular PTZ tour mode is

maintained, but at night the camera

points toward a sensitive perimeter

line or a high-risk grati spot.


Always Know Who You’re Talking To.™

Tracking algorithms are used to followeach classified “blob” as it moves throughthe field of view.



32 urgent communications august 09

Given the advances o server

processing power, a central

analytics processor implementa-

tion provides clear added fexibility

compared to an edge implementa-

tion, while costs o implementation

between a central processor–based

and edge processor–based deploy-

ment are nearly equal. A central

video analytics processor is there-

ore recommended. In terms o costs,

or a 12-camera implementation, the

cost per camera roughly would be in

the range o $3,000 to $6,000, assum-

ing a standalone project adding video

analytics to an existing surveillance

system on a turnkey basis.

In terms o application, there

currently is not an all encompass-

ing analytics solution that can be

expected to alert any suspicious

behavior in a citywide deployment

model. Available packages, how-

ever, do oer a airly broad range o

behaviors, based on classiication

o people, vehicles, etc. Also, most

video analytics companies are very

open to working with clients to

add behaviors and unctionality to

their packages, in response to spe-

ciic needs. Based on the growing

size o surveillance systems within

cities and the associated need or

intelligent video, a ast evolution

o analytics packages is expected

with possibly a proound impact on

the unctionality and operation o

these video surveillance systems.

Already today, analytics packages

should be able to provide beneits

or speciic applications, and the

timing is right or cities to start

testing and implementing analytics,

at least on a small scale or specic

uses. Particularly, with a central

processor–based approach, imple-

mentation is simple, and a specic

analytics license can be transerred

rom one camera to another. This

allows cities to test analytics in a

pilot setup at dierent locations in

a network, without spending a large

budget right away. As cities take

their rst steps into video analytics,

by orming partnerships with their

integration partners and analytics

vendors, video analytics might soon

develop into a mandatory part o

surveillance systems rom day one

o deployment and undamentally

change the mode o operation. n

Leonhard Korowajczuk is the chief

executive officer of CelPlan Technologies.

Jasper Bruinzeel is the company’s vice

president of marketing and sales.


analytics article aug09

Documents