WHEN you talk to
suppliers of analogue, standard IP and even HD cameras, there is often the
suggestion that megapixel cameras come with baggage so significant it impinges
on their ability to handle real world applications. But such talk ignores the
strengths of megapixel camera technology.
One of the
beautiful things about megapixel cameras is that they capture an entire scene
all the time and to great depth. That means with a megapixel cameras you’re
installing a virtual PTZ that retains the ability to zoom into a vast
horizontal and vertical scape long after the footage is recorded. In
comparison, ordinary PTZs record images based on the focal length and the field
of view at any given moment of a timeline.
Megapixel cameras
are so good it’s hard to deny every external system should have some as part of
an overall solution. Sure, you can drone on about challenges like image stream
size and the need for adequate light levels with MP, but the benefits of a
topical MP application are simply undeniable and their power makes them worth
the spend.
Another perceived
negative that invariably pops up when talking MP is the need for capable and
expensive system elements in order to supply the necessary bandwidth and
storage. But once again, such talk ignores the payoffs in improved performance.
While you may need
to option shared network to handle large MP camera numbers and resolutions, this
is not fundamentally challenging and the gains outweigh the costs in a number
of areas and the reduction in camera numbers will make the system simpler than
it would be with analogue or standard IP cameras.
A major benefit of
megapixel cameras is their ability to replace multiple standard IP or analogue
cameras. It’s a fundamental that means the math is not one for one – instead
it’s one for 3 or 4 and sometimes many more. Such elements need to be taken
into account when it comes time to make a decision.
Balancing act
There are 2 major
suppliers of megapixel cameras in Australia – Arecont (distributed by Pacific
Communications) and Avigilon. Each has taken a different approach to the
challenges posed by megapixel technology and to its delivery and storage on
networks.
Avigilon employs
scientific grade CCD cameras, JPEG2000 compression and clever software in order
to deliver huge dollops of performance. Avigilon’s High Definition Steam
Management is the key to its superb performance on networks. The company’s
power Control Center software also plays a role and offers excellent features
like auto lighting control and auto contrast controls.
During a product
demonstration the company’s Matthew Fisher shows me how Avigilon’s purpose-built
proprietary Control Center software handles multiple MP cameras with ease
thanks to HDSM.
“The choice of
Video Management System (VMS) is the key when working with larger MP cameras,”
Fisher explains. “Without a purpose-built VMS, the system may not be able to
fully handle the images generated by large megapixel cameras and introduce
additional installation challenges.”
Fisher says there
are ways to ensure a network is not overloaded with unnecessary data when using
megapixel technology.
“For example, when
you drill into a scene we only send those image packets that are required to
view that part of the scene – we don’t send all the superfluous information,”
he explains.
“Using the HDSM capability
of our Control Center software we can view 3, 5, 11MP cameras even across a
simple ADSL connection. And this efficiency applies at the workstation level. We
can run 6 x 16MP Avigilon megapixel cameras on a laptop and have multiple other
windows open with multiple zoom levels. It works because our system does not
load up the chipset with huge levels of processing.”
“Using out HDSM you
can run 6 x 16MP Avigilon megapixel cameras on a laptop and have multiple other
windows open with multiple zoom levels. It works because our system does not
load up the chipset with huge levels of processing”
Matthew Fisher,
Avigilon
As part of the
demonstration we take a look at a 16 megapixel camera view of a large open area
in a city including a tramline, a dock area, a large carpark, a public square
and various buildings and carparks. The scene is huge yet this one camera is
viewing and recording all of it.
It’s when we
start zooming in that the fireworks start. We can see number of trams, we can
read the names on the boats, we can identify people throughout the entire scene
(it’s a huge view, remember) and we can identify car makes inside a nearby carpark.
When we look at
the street we can see number plates. There’s nothing in the scene that we can’t
zoom in on. It’s awesome performance and when Fisher points out that we’re
looking at a camera capable of replacing 50 analogue VGA cameras, I look at the
scene and think – yes – that seems right.
Such capability
is where the strength and economy of megapixel cameras becomes obvious.
Installing dozens of surveillance cameras is a seriously expensive business,
whereas installing a single megapixel camera at the end of a single blue cable
is a no-brainer.
Using Avigilon’s Control
Center and taking advantage of Avigilon’s HDSM, Fisher starts calling up
multiple camera views and the laptop all this is running on doesn’t miss a
beat. The ability to pull up windows within windows and zoom in on them while
still viewing the whole scene shows the great power of the Avigilon solution.
You can see the system would be ideal for video walls thanks to its ability to deliver
what Fisher call ‘situational awareness’.
Best of all,
performance remains superb – there’s no boxing inside images – no sign of a
compression engine running on the edge of its capabilities.
“Because we are a
medical imaging company we learned early what the downfalls were with any
interframe compression codec very early on,” Fisher explains.
“As a result we
use JPEG 2000 as our compression codec – our experience means we know what the
limitations of the other compression algorithms are. With other codecs you are
very limited to what you can do. They are lossy, they have a high error rate
and high latency.
“They have to
rebuild the video streams. You see scenes in which artefacts from previous
images remain in the current image all because of the challenges of constantly
having to rebuild the video streams.
“With our
JPEG2000 we can go through streams frame by frame but with an interframe
compression codec you have to go back to the last I-frame and then step
forward.
“This impacts on
your ability to drag a timeline. On our Control Center NVMS or viewer you can
just grab a timeline and pull it backwards or forwards and everything updates
instantaneously and beautifully because there is no the high processing need on
the decoding side of the client.”
According to
Fisher, with interframe compression codecs there are issues with the level of
movement in a scene and how it impacts on the average bit-rate of a camera.
“At 80 per cent
movement the camera will go to full I-frame anyway – so is that actually saving
bandwidth? It is when there’s low movement but the saving comes at the expense
of image quality.”
As part of the
demo we take a look at the image from a very good analogue camera installed
beside the Avigilon unit. Fact is, it’s not possible to tell the some license
plates with absolute certainty on the analogue camera. Even the colour
reproduction is not as good as the Avigilon megapixel – it’s interesting to see
how capable by Avigilon’s cameras really are.
Next we check out
a single Avigilon 5 MP cameras inside in food court – we can zoom down to the
lifts, the scene lightens and we can see the what’s going on down there – we
make a window and use that part of the scene as a camera view.
“You’d need half
a dozen cameras to do the same job as the 5MP and you would not have the
ability to drill into scenes,” Fisher explains. “It’s possible to clearly see
the possible savings of using a single Avigilon camera in open areas.”
With a reputation
as a manufacturer of 16MP cameras, it’s possible to forget that Avigilon
actually has a wide range of cameras including 1, 2, 3, 5, 11 and 16MP units.
An obvious question that springs to mind when considering the depth of this
range is which cameras to use where.
“The question of
what cameras to use and what quality of images it’s possible to get from scenes
relates to a number of factors – the angle of the lens, the distance to target
– it’s a big question,” explains Fisher.
“Avigilon has a
clever software tool to assist with lens and camera choice – you plug in a mounting
height (say 4m), distance to target (say, 15m), scene width (say 12m). Once
this information is entered into the system, the integrator knows there’s a
42-degree angle of view and the system has also advised the installer of the
required focal length in mm needed to achieve that scene and then the pixels
per metre – that’s the resolution,” he says.
“The resolution
might be from say low – 67 pixels per metre up to high resolution photographic
quality of 328 pixels per metre in the scene. You then need to choose what sort
of images you want. Do you want full facial recognition throughout a scene? Do
you just want a general surveillance application? Or do you want to be able to
recognise clothing colour, gender, general demeanour and what the person is
doing?”
As Fisher quickly
demonstrates, once all the parameters are plugged in it’s possible to establish
that a 1MP camera of 720p will give a pixel count of 107 pixels per metre
viewing this scene. With 2MP it’s 160 pixels, while 3MP gives 171 pixels, 5MP
gives 216. 11MP offers 334 and 16MP dishes up a stonking 406 pixels per metre
of resolution. To put this into perspective, around 165 pixels per metre is
considered a high quality image.
“Using this tool,
the system allows you to establish everything you need from lens to camera
angles or view – horizontal and vertical, scene resolution, combined
camera/lens length,” explains Fisher. “Each camera view can be engineered in
this way. Across a scene you might want general surveillance and at choke
points you might want face recognition and license plate recognition.
“This allows you
to choose applications when you would need to use a 5MP or a 16MP camera. I
would argue there really is not a niche for large megapixel cameras – the scene
you need to cover determines the camera you need. Why put in 6 2MP when you
could put in a single 16MP in applications like stadiums, airport lobbies,
loading docks – locations you might not be able to get around and cable multiple
cameras – there might only be a single position but need to cover the entire
scene.”
H.264 is the key
At Arecont
Vision, Raul Calderon agrees that megapixel cameras are suited to all
applications and should not be limited to just external or very large indoor
applications.
“Multi-megapixel
cameras using H.264 compression from Arecont Vision are suited for all
applications,” says Calderon. “It really depends on how megapixel cameras are
implemented. The important thing is to understand whether you want increased
resolution, decreased camera count relative to standard definition cameras or a
combination of higher definition and camera reduction.
“Megapixel
cameras are exceptional for installations that require large areas of coverage
or extreme detail for forensic purposes. Frankly, they deliver the image
quality that one has always wanted but rarely ever achieved out of their video
security system.
“For example, a
recent installation employing Arecont Vision MegaVideo Day/Night dual-sensor
cameras in a newly constructed traffic tunnel in South Korea was implemented to
capture license plates at high speeds in a dimly lit environment – a feat that
would prove very challenging using conventional cameras considering the pixels
on target requirement to capture license plates,” he explains.
“Another example
is the installation at an American high school where coverage was needed for an
open, hexagonally shaped hallway. The imaging solution incorporates Arecont
Vision’s 5 megapixel cameras with ultra wide angle lenses to provide warp-free
images of high traffic areas with the ability to electronically zoom in live as
well as after event to extract forensic information and view the smallest of
details in high definition of these applications clearly refute the
misconception that multi-megapixel cameras are for specialty applications.
“The truth is
that with H.264 compression, bandwidth is no longer an issue making
multi-megapixel cameras ideal for virtually any application as they can greatly
reduce the actual number of cameras needed to cover an area versus conventional
cameras, and lower the total cost of ownership for customers.”
And Calderon says
that while there are always challenges with network bandwidth and storage H.264
has changed things completely.
“H.264
compression has alleviated the bandwidth issues associated with early JPEG
based megapixel cameras,” he explains. “As a result, the amount of storage
required is significantly less than it used to be. Calculating the proper field
of view to achieve the pixel density desired within the scene can also be a
challenge.
“To assist
installers with these challenges, Arecont Vision provides calculators to help
in designing a system with the right amount of storage as well as the correct
field of view and pixel density within the scene to resolve the target.”
Calderon says
that another commonplace challenge for security professionals is to understand the
long-term value and return on investment (ROI) that large format megapixel
cameras offer relative to cheaper and far less efficient standard definition
surveillance cameras.
“Translating the
technical capabilities and benefits of this relatively new technology into
management and bottom-line benefits can be difficult as it is challenging to
assess the cost-savings these devices provide relative to what they can deter
or prevent from occurring,” he explains.
“Fortunately
multi-megapixel cameras are proving to be more efficient in large system
configurations as they actually reduce the physical number of cameras required
to cover large areas versus conventional cameras, and consequently reduce the
associated installation expenses.
“The latter
provides a tangible measure of cost-efficiency. But the superior performance
capabilities of multi-megapixel cameras to help ensure cost avoidance and
improved forensic information are not readily documented at this early stage of
multi-megapixel cameras’ deployment history.”
According to
Calderon, big megapixel cameras are not expensive.
“Consider Arecont
Vision’s new AV10005 10MP camera which provides two times the pixels but only
costs 10 per cent more than its AV5105 5MP camera,” he says.
“Multi-megapixel
IP cameras and systems can make a profound positive impact on budgets because
they vastly improve image quality and improve ROI. For example, a singular 180
degree panoramic 8 megapixel camera (AV8185) from Arecont Vision can been used
for parking lot surveillance replacing as many as 10 to 15 analogue cameras,”
he says.
“Not only is
there a reduction in total installation costs but there is also a reduction in
total storage requirement and fewer cameras to manage, service and maintain. This
ROI case is made time and time again all over the globe for virtually any
megapixel parking lot or wide open area application.”
As Calderon
explains, while the actual camera device might be a little more expensive, the
total solution cost or total cost of ownership is not.
“Megapixel IP
cameras on IP recording platforms are extremely cost effective,” he says. “It
takes fewer megapixel cameras to cover the same physical area and picture
quality are greatly improved over standard resolution cameras.
“And using fewer
cameras offers the additional benefits of fewer cables, fewer camera housings,
fewer lenses, fewer network video recorder (NVR) licensing fees, fewer NVR
servers, and far less installation expenditures.”
In general,
higher resolution cameras have increased performance challenges in low light
environments and Calderon says Arecont Vision has implemented several
technologies to combat this issue.
“For example, our
new AV10005 is the world’s first 10 megapixel and 1080p H.264 Dual Mode camera
that resolves many of the issues commonly associated with multi-megapixel
cameras,” he says.
“The camera
offers a binned mode that combines and averages pixels for better low-light
performance, greater sensitivity and improved signal-to-noise ratio (SNR).
“Other features
include on-camera real-time motion detection with our enhanced noise
cancelation technology, a programmable shutter speed to control motion blur,
and MoonLight mode for extended exposure. The performance and functionality of
this camera will change the landscape for multi-megapixel cameras.”
And Calderon says
that in the future larger megapixel cameras will be commonplace.
“We firmly
believe that megapixel cameras will eventually become the dominant platform for
general video surveillance applications we are already seeing evidence of this
as megapixel popularity continues to grow,” he says.
“Video surveillance
imaging, recording and processing technologies continue to evolve rapidly and
driving the demand for higher performance systems. As more installations employ
megapixel cameras with high levels of success, the industry’s migration to
megapixel cameras will continue to accelerate.”
“The truth is
that with H.264 compression, bandwidth is no longer an issue making
multi-megapixel cameras ideal for virtually any application as they can greatly
reduce the actual number of cameras needed to cover an area versus conventional
cameras, and lower the total cost of ownership for customers”
Raul Calderon,
Arecont Vision
