CAMERA choice has
become much harder – not just in terms of selecting between camera technologies,
but trying to establish which of the multiple selections allows the best
balance of resolution and performance. When thinking performance, think frame
rate across a network, as well as bandwidth and storage demands.
The camera
technologies now available to the security manager or integrator include
analogue, IP (CMOS and CCD), hybrid analogue and IP, HD (720 and 1080), and
megapixel. It’s this last genera that we’re going to focus on here and the
reason for this is that megapixel is most likely to be the gold standard of the
future. It’s also the most inherently variable. A megapixel camera can be 1.3
Megapixels or 15 Megapixels, or anything in between.
There’s a reason
megapixel technology has become popular over the past couple of years. For a
start all megapixel cameras are IP. But most importantly, megapixel offers
improved resolution. A 1.3 megapixel camera has 4 times better resolution than
the 640 x 480 line resolution of standard VGA.
This higher
resolution gives users a number of benefits. It can be traded, allowing a
single camera to be used to view a larger area. Alternatively, it can be broken
up so as to allow multiple scenes to be recorded or viewed from a single
camera. Megapixel also offers benefits like improved recording of fast moving
objects, better resistance to blooming, and low power sensors that support PoE
applications.
There are a
couple of other important considerations with megapixel cameras. For a start,
many cram 3 times the usual pixel numbers on a standard 1/3-inch chip and to do
so pixels are shrunk with a corresponding loss of sensitivity in low light.
Some megapixel cameras are better than others but keep an eye out for
AGC-generated noise – especially if you’re recording at high resolutions and
frame rates. Look for ½-inch chips for your low light apps if appropriate
lenses are available.
Another thing
you’ll need to take into account is the fact you must use megapixel lenses on
megapixel cameras or performance will be eroded. And think about frame rate. It
will be lower than the 25 images per second you get with VGA, D1 and HD.
Megapixel frame rates are typically between 8 and 15 images per second. It goes
without saying that quality and frame rates are a juggling act requiring
innovative system designs incorporating local/onboard storage.
Know what you
need
According to
Bosch’s Sean Borg, megapixel technology refers to CMOS chip sensors that have
the ability to resolve very high resolution images under certain light
conditions.
“A buyer should
understand exactly what their real requirement is as far as recognition or
detection is concerned, and to what level, before they buy megapixel,” Borg
explains. “Do they want to know there is someone there? Do they want some
descriptive features of a person – to be able to recognise the person in a
court of law?
“The answers to
these questions combined with the limitations of camera mounting and
positioning options allow you to make a more informed decision about which type
of megapixel camera you will need,” Borg says.
“Other things to
consider will be the speed of the object you expect to capture and the light
levels within the area of capture, this will also assist the decision. The last
major consideration will be the network pipe size you are prepared to have and
the storage requirements.”
According to
Borg, megapixel cameras should always be transmitted using JPEG or MJPEG
compression, otherwise it’s better to save your money and buy a nice IP camera.
“A megapixel
camera will stream good quality video at around 30Mbit’s per second or higher,”
he explains. “To then strip that down to MPEG4 or H.264 and stream it at say,
1.2Mbps, means you are wasting most of the dollars you invested in the
resolution aspect to achieve something a good IP camera can easily produce – this
does not make a lot of sense.”
“There’s an
obvious trade-off between resolution, frame rate and storage capacity with
megapixel IP. I believe the key to that trade-off is management of video using
profiles and analytics to help manage the storage and bandwidth issues”
Sean Borg, Bosch
Security Systems
According to Borg
the greatest strength of megapixel technology is great resolution in good
lighting.
“There’s an
obvious trade-off between resolution, frame rate and storage capacity with
megapixel IP,” he explains. “I believe the key to that trade-off is management
of video using profiles and analytics to help manage the storage and bandwidth
issues.
“The sweet spot
would be having a system that only transmitted and recorded nice, raw JPEG
images at the time of the incident, clearly that would be the ultimate goal but
it’s not always possible, especially for a surveillance application rather than
a security application.”
According to
Borg, when it comes to megapixel and HD, he believes, whichever of the 2 technologies
offers the best image quality in varied light will dominate.
“The price of
technology will always reduce to an acceptable level,” Borg says. “This means
image quality will be the determining feature.”
At Mobotix,
Graham Wheeler says that when buying megapixel cameras buyers should consider
megapixel camera as having more than one million pixels which is the number of
individual recognition points on the sensor.
“Most analogue
CCTV cameras record at resolutions between CIF (0.1 Megapixel) and VGA
(approximately 0.3 Megapixel), this compared to 1 Megapixel being between 10
and 3 times the resolution,” Wheeler says.
Wheeler says
resolution is arguably the greatest benefit of megapixel technology.
“Generally, the
biggest complaint about CCTV cameras is “why is the image quality so poor?”
Wheeler explains. “The answer is that this is due to low resolution sensors and
analogue systems. In comparison, our products have 3 Megapixel colour sensors
which give up to 30 times the resolution of a CCTV CIF recording.
“What this means
is that you can use fewer cameras to cover the same area,” he says. “I was
recently on a course in Perth with our integration partners and was shown an
image from one of our 3 Megapixel Q24 cameras that replaced 16 analogue cameras
mounted on 2 different poles to give an overview of a playground. So if
megapixel is deployed correctly it can not only provide a better result but
also save you money.”
Wheeler says
backlight compensation is excellent with the latest MOBOTIX CMOS sensors.
“We can set
regions in the field of view as the reference light level so that we can
compensate perfectly for high back light areas. This makes our cameras ideal
for the entrance to a bank, shop or in a mobile application such as a bus that
has major back light issues,” he says.
“Having no iris
also helps with this as you actually shutter the sensor pixels rather than
change the iris.”
In Wheeler’s
opinion, the current sweet spot with megapixel technology is between 1 and 3
megapixels but he says this will increase in years to come.
“Decentralised
systems, which is what MOBOTIX has pioneered in this market for a decade works
exactly the way computer networks work. This is exactly how computer networks
are built together with computers on networks and they back up to NAS or
storage servers when the network is available”
Graham Wheeler,
MOBOTIX
“Currently most
people record 0.1 to 0.3 Megapixel, so to go up in resolution to 3 Megapixel is
between 10 and 30 times increase in storage demands. I think that is a good
improvement to start with and looking at the current costs on storage you can
realistically use these sizes of images and record them if required for a
month.”
Wheeler concedes
low light can be a challenge for megapixel cameras that have too many pixels on
their sensors.
Low light
performance
“MOBOTIX uses a
1.3 megapixel black and white sensor in its cameras to get around the low light
challenges,” Wheeler says. “We all know that B/W sensors give better low light
characteristics and that is because of no IR cut filter and fewer sensor
points. Consider that the piece of silicon that makes up a colour sensor has 3
times as many sensor points because you have to sense red green and blue for
each pixel point to make up the colour image.
“In the same size
piece of silicon a monochrome camera has 3 times less sensor points as it only
senses black and white. This means that the sensors can be 3 times bigger, so
they are 3 times as responsive in low light. So what we chose is a large sensor
with 1.3 Megapixel on the sensor that gives excellent low light characteristics
against a 3 megapixel sensor that has up to 9 times more sensor reference
points in the same area.
“It’s all about
the light falling on the sensor point, the more light the better the low light
response,” explains Wheeler. “Many larger camera pixel camera manufactures
produce a great image in full daylight but a very poor image at night, much the
same as your camera on your phone.
“So in low light
conditions, a higher pixel count does not necessarily mean better quality. In
time, the black and white sensors will improve and we will increase the pixel
count to 5, 8 and so on, but at the moment 1.3 to 3.2 Megapixel is the sweet
spot.”
Wheeler says he
has heard some manufactures of cameras argue that the industry will never need
more than HD resolution because our HD TV’s look great.
“But with every
HD movie you see at home there is a production crew behind every camera zooming
the camera to make sure you see the exact image they want you to see at just
the right time,” Wheeler explains.
“In CCTV
applications we have to put one camera up to cover a large area, then after the
event security staff are the production crew and zoom in to the recorded image
to get the needed image afterwards causing us to need much higher resolution
than TV.”
Wheeler makes the
interesting point that in the future the video surveillance industry will use
much higher resolution cameras than broadcast TV.
“Ultimately you
could consider using a single 20-megapixel camera to cover the front of an
airport and then afterwards zoom into the recorded image to get a person in the
crowd where you need 20 cameras using current technology.”
Wheeler says that
a key element of the Mobotix solution is that it is decentralised, not
centralised.
“Centralised
systems are exactly how our old analogue CCTV systems were designed 40 years
ago,” he explains. “They were built with a central DVR with cameras sending
data to the DVR to do motion detection, etc. If you do the same with an IP
camera this is just how a computer network operated 20 years ago with a dumb
terminal on a network and a central mainframe computer.
“All the data is
sent regardless of whether it is needed to the central server or multiple
server for motion detection. Consider Singapore Polytechnic with over 1000
megapixel cameras in it all running motion detection. You would need about 250
to 500 PC’s centrally running software to manage it. If the network goes down
you lose your footage, if your servers fails you lose your video footage –
that’s 2 points of failure.
“Decentralised
systems, which is what MOBOTIX has pioneered in this market for a decade work
exactly the way computer networks work,” Wheeler says.
“Each camera has
a processor in it that handles the logic and storage buffer that can bridge
network and server/NAS failures. This is exactly how computer networks are
built together with computers on networks and they back up to NAS or storage
servers when the network is available.
“In this
structure if the network fails, the cameras buffer the image so no data is lost,
if the storage server is lost then the camera buffers the data so again, no
data is lost. This system has no single point of failure and is a vast
improvement over analogue CCTV systems designed 60 years ago, as well as a vast
improvement over centralized systems that used the computer network design of
20 years ago.”
Wheeler says the
key thing is the decentralized approach saves system costs by removing the need
for costly centralized servers.
“In Singapore
Polytechnic 1000 MOBOTIX cameras all record through just 2 servers to NAS
devices rather than the 500 or so PCs that would have needed a storage area
network in a centralized system” he says. “The issue with IP solutions is that you
have to compare system costs, not unit costs.”
Mobotix’s latest
cameras are a complete camera and DVR/NVR in a waterproof housing.
“There’s USB and
SD card for long term storage at the camera, CMOS sensor for strong backlight
compensation, no moving parts – no iris or IR cut filter – and PoE in all
weather conditions thanks to an integrated heater.
“There’s also
MxCC Control Centre Viewing interface, camera lens, housing, VoIP, redundant
architecture, integrated video motion detection, a PIR, temperature sensor and
audio triggers all onboard. It’s not comparable to an analogue camera that it
simply a signal source.”
HD or megapixel?
At Axis
Communications, Wai King Wong says that megapixel network cameras “do not
adhere to any standards, but rather are an adaptation of the industry’s best
practices and refer specifically to the number of image sensor elements of the
digital camera.”
He says megapixel
and HDTV technology enables network cameras to provide higher resolution of
video images than analogue CCTV, giving the ability to see details and to
identify people and objects.
“This is a key
consideration in video surveillance applications. With a megapixel or HDTV
network camera, the resolution is at least three times better than with an
analogue CCTV camera,” Wong says.
“Megapixel and
HDTV network cameras have an important role to play in video surveillance
applications. They are obviously more suitable for some areas than for others.
The determining factors are ultimately the requirements of the individual
customer’s system. Once the specific objectives for each network camera in the
surveillance application are set, the right type of camera can be selected for
the best fit.”
According to Wong,
the comparative strength of the new technologies boils down to a simple truth –
megapixel and HDTV mean higher resolutions giving more detail – that means
better identification.
“Megapixel
network cameras are excellent for many applications, for example when video
with extreme image detail is required, such as in overview surveillance of
banks, transportation hubs and other premises,” Wong explains.
“In addition,
some megapixel cameras offer “multi-view streaming”, which means that the same
camera can deliver different video streams from different areas of a scene,
which essentially leads to one megapixel camera replacing several standard
cameras – for example, by monitoring more than one cashier in a retail store.
Digital pan/tilt/zoom is another advantage offered by megapixel cameras.”
Wong says that in
the surveillance industry, some best practices have emerged regarding the
number of pixels required for quality images in certain applications.
“We see the
future of video surveillance as industry standard HDTV (SMPTE compliance) and
H.264 video compression. Megapixel provides high quality images but getting a
5-megapixel camera to run over an Ethernet network at a high frame rate
requires a lot of processing power, bandwidth and storage”
Wai King Wong,
Axis Communications
“For an overview
image, it is generally considered that 20 to 30 pixels are enough to represent
one foot of a scene,” he explains. “For applications that require detailed
images, such as face identification, the demands can rise to as much as 150
pixels per foot. This means, for example, that you want to be able to strongly
identify people passing through an area that is seven feet wide and seven feet
high, the camera needs to provide a resolution of 1,050 x 1,050 pixels, which
is slightly more than 1 megapixel.”
When comparing
megapixel to HD, Wong says a true HDTV network camera complies with industry
standards which ensure excellent colour representation, full frame rate and a 16:9
format.
“HDTV gives full
frame rate video and excellent colour representation,” says Wong. “It provides
a wide-screen 16 x 9 image, perfect for modern screens and TVs. The fact that
true HDTV network cameras comply with key industry standards means that image
quality is guaranteed to a higher degree than for many megapixel network
cameras. HDTV network video is superb for applications where full frame rate is
prioritised.
“Meanwhile, two
drawbacks of regular megapixel technology include the lack of colour quality as
compared to the human eye, and the fact that at higher megapixels you don’t get
full frame rates. For example, a casino may use chips that are pink and red. A
megapixel camera will be challenged to differentiate between the two. In
addition, if you need fluid frame rates, megapixel cameras may only offer 5fps
or 10fps at higher resolutions.
“In such cases,
you need to use HDTV cameras, which offer full 30 fps, colour representation
that matches the human eye and are based on standards (SMPTE),” Wong says. “An
additional benefit of HDTV cameras is being able to use a 16:9 ratio versus a
4:3 ratio.
“The benefit of
the wider ratio (16:9) is that you can make sure your pixels aren’t wasted on
ceilings, sky or ground, but the viewing area you want to see.”
According to
Wong, there are two factors in determining the pixel count in a scene: the
amount of pixels in the image sensor and the lens type used.
“When aiming for
a broad viewing area, you need to use a wider angled lens,” he says. “With this
wider view you dilute the amount of pixels available in your target area and
limit the amount of detail seen after the fact. When monitoring a parking area
without a need for license plate information, this won’t matter.
“On the other
hand, if you need to identify a person or a license plate, select a more
narrow-angle lens to obtain high detail. Thus, narrow lenses are used more for
identification and after-the-fact forensics.”
Wong says there
are four main considerations when looking at the sensor type needed.
“First decide
whether you are going to use the megapixel network camera for after-the-fact
forensics,” he says. “Next, given your specific application, decide the number
of pixels required and then decide whether you require full-colour representation
of images or high frame rates. Finally decide whether you want multiple views
from a single camera.”
Wong says it’s
important that when determining the right megapixel count, users know ahead of
time what detail they will want to zoom in on.
“For example,
suppose you are setting up cameras for a teller area and your target pixel
count for facial recognition is 60 pixels from ear to ear to identify a person
(60 pixels is a UK requirement for admission in court),” Wong explains.
“Measure the
distance from the camera to bank customers to mathematically determine the
required megapixel count. The same goes for identifying license plates. If you
know the camera is 100m from the viewed location and need 15 pixels for the
height of the license to get character recognition, then you can mathematically
determine the sensor required. Online tools are also available to help
determine the right megapixel sensor and lens type.
“You also need to
determine if your customer wants to replace multiple fixed VGA cameras with one
megapixel network camera,” he says. “A practical example is in a retail outlet
in which you want to cover three cash registers with one network camera. To do
so, you’ll need to find one that can record multiple streams from the same
sensor at the same time. Not all megapixel network cameras are capable of this.
“Choosing a lens
is the final factor. The more optical zoom incorporated, the fewer pixels
required on your sensor. The trade-off is a smaller viewing area. The contrary
would be true for wider angle lenses.
“To assess which
network cameras you need (megapixel, HDTV and/or non-megapixel, including
pan/tilt/zoom cameras) it is important not only to do the calculations, but
also to survey the location to determine the number of interest areas, the size
of these areas and whether they are located close to each other or spread far
apart,” he explains.
“Other
considerations should also be taken into account; for example the availability
of guards performing live monitoring, the need for light sensitivity, bandwidth
and storage.”
Wong does not
agree the future of video surveillance will be megapixel IP incorporating local
storage and PoE.
“No, I don’t
agree with this” he says. “We have megapixel cameras up to 3 megapixels in size
but we see the future of video surveillance as industry standard HDTV (SMPTE
compliance) and H.264 video compression.
“Megapixel
provides high quality images but getting a 5-megapixel camera to run over an
Ethernet network at a high frame rate requires a lot of processing power,
bandwidth and storage – although this can be mitigated by using MPEG-4 or H.264
compression.
“That means you
must be careful to match the appropriate resolution to your application,” Wong
says. “Luckily, determining your requirements is linear and easy to figure out
once you know your objectives. If bandwidth and storage limits were no factor,
you’d probably choose the most pixels available, but even a 1.3 megapixel
network camera has four times the storage and bandwidth requirements of a
standard VGA camera.”
Broadest
megapixel range
Meanwhile Arecont
Vision’s regional manager, Ethan Maxon, says Arecont Vision offers the broadest
range of megapixel cameras on the market, from 1.3MP up to 10MP, all using
H.264 compression, and available in MegaDome all-in-one dome models and
dual-sensor day/night versions.
“Our Surround
Video Panoramic IP cameras incorporate four 2MP image sensors (total of 8 MP)
to provide either 180-degree or 360-degree high-definition views to do the job
of multiple cameras while providing high resolution coverage of large areas.
“Day/night
versions of Arecont Vision’s cameras have single sensors with IR cut filter or
dual (color and B/W) sensors to ensure even better low-light sensitivity,” says
Maxon. “Switching between a 3MP color image sensor for daylight and a separate
1.3MP black-and-white sensor for lower light levels ensures optimum image
quality regardless of available light levels.
“H.264
compression has been an industry game-changer by making megapixel technology
even more deployable. While retaining the same level of image quality as Motion
JPEG, Arecont Vision has been able to dramatically decrease video bandwidth
levels.
“Finally, in
Australia, we have found a software agnostic approach to be a dramatic differentiator.
Network Video Recorder (NVR) vendors such as Exacq, DVTel, Verint, Genetec,
Milestone, and 80 others, all have a strong presence and Arecont Vision is
compatible with all of them.”
Maxon says
comparing an analogue camera to a megapixel camera is like comparing HDTV to
standard definition TV.
“In the camera
world, for example, a 1.3-megapixel (1.3MP) camera employs a 1280 x 1024-pixel
image sensor with a total of 1,310,720 pixels,” he says.
“A
conventional-resolution image (comparable to the analog PAL format) would
provide 704 x 576 pixels or only about 400,000 pixels. Increasing the pixel
count or pixel density within a sensor increases the resolution – a 3MP image
has more than 3 million pixels and a 5MP image more than 5 million, and so on.
“It’s the
additional pixels in a megapixel image that provide so much more detail both in
real-time and in playback mode, greatly aiding in forensic investigations as a
result of the density of data captured. Operators can observe an entire field
of view while simultaneously zooming in for tight detailed views.
“Even stored
images can be zoomed in after event to view specific parts of a larger image
with extreme detail,” Maxon says. “When you consider the concept of
“pixels per metre” (that a certain number of pixels is required to
depict one meter of a scene for a specific application), it’s clear that more
pixels equates to an ability to closely view much larger areas.”
Maxon says that
using fewer cameras to cover large areas translates into cost savings related
to infrastructure (cables, mounts, housings, etc.), which makes it easy to
realize a positive return-on-investment (ROI).
“In Australia and
in other countries, we have found our greatest strength has been the ability to
deliver high image quality at very high frame rates while minimizing bandwidth
usage with H.264 video compression,” he explains.
“For example,
just days after installation, a major Australian retail chain was able to
finally identify a burglar who for months had been stealing laptop computers
from multiple store locations.”
Given the
trade-off between resolution, frame rate and storage capacity with megapixel
IP, Maxon says selecting the right megapixel camera needs to be closely aligned
with the specific application and the customer’s expectations for performance.
“In general,
Arecont Vision’s H.264 2MP camera achieves a great balance of image quality,
low light sensitivity, frame rate, low bandwidth usage, and low storage
requirements,” he says.
“Specific to
Australia, we have had a lot of success with Arecont Vision’s 180-degree and
360-degree 8MP megapixel panoramic cameras in retail chains throughout the
country. The ability to capture a “situational” view of store areas displaying
valuable merchandise has turned out to be a huge competitive advantage for our
system integration partners.”
Maxon sees
megapixel and HD as having their own applications but he says megapixel offers
superior performance.
“The HDTV
standard represents a subset of possible megapixel images,” he explains. “HDTV
images are classified as 720 lines (0.9MP) or 1080 lines (2.1MP). Arecont
Vision supplies cameras providing resolutions from 1.3MP up to 10MP, and any of
the cameras can be configured to customizable cropped resolutions less than
their original native resolution, and thus are comparable with HD.
“The HDTV format
and frame rate are appropriate for some applications, but Arecont Vision
cameras provide a range of resolutions (including standard HDTV resolutions),
frame rates and multi-stream compression to provide better image quality,
broader field-of-view, and greater flexibility for the system designer and
user.”
When it comes to
making selections of megapixel IP cameras, Maxon says that megapixel imaging is
useful to any application and represents a significant upgrade in system
functionality compared to conventional analog and VGA imaging devices.
“There are very
few reasons to buy analog or VGA cameras anymore – they are inferior to
megapixel cameras and do not provide an upgrade path,” he explains. “Customers
also need to consider storage resources, network bandwidth, lighting, lenses,
and other environmental and situational factors.
“The professional
security industry is beginning to realize that, in addition to lower bandwidth
storage requirements, megapixel cameras can dramatically decrease costs related
to other elements of a system such as fewer software licenses, a reduction in
cabling, fewer lenses, and a decrease in the installation time needed to bring
it all together. Much of the potential savings comes from being able to reduce
the number of cameras needed on a specific project.
“For instance,
where 10 cameras might have been required to cover a parking lot this might now
be accomplished with 3 or 4 strategically positioned 3 megapixel cameras or
just a single panoramic camera, depending on the application requirements,”
Maxon explains.
“To compare just
the price of a megapixel camera with an analog or standard IP camera does not
tell the story taken out of a system context. To compare apples to apples, you
need to cost out and compare the total number of cameras along with the total
infrastructure needed to support them. Then the efficiencies of megapixel
cameras are clear to see. Today, Arecont Vision’s compact JPEG camera is comparable
in price to an analog camera with an encoder.”
According to
Maxon, the future is going to see features like power-over-Ethernet as a part
of every IP camera in the foreseeable future; a fundamental already true for
the majority of IP cameras. But he does not see storage at the camera in every
application.
“I don’t think
everyday applications require local SD storage. With the efficiency of today’s
leading video management systems employing push video and with more advanced
NVR based systems managing live video, stored video and playback,” he explains.
“Local SD
provides little benefit as it has low-capacity storage and has to be
over-written. Multi-streaming capabilities of cameras allows recording on
multiple computers, addressing security concerns typically invoked as
justification for SD storage.”
According to
Maxon, Arecont Vision’s product range has sensor resolutions from 1 to 10MP for
its single-sensor box cameras – that makes it’s range among the most complete
on the market.
“Arecont Vision
is able to achieve frame rates of approximately 7fps at 10MP (3648 x 2752),
which is faster than many manufacturers’ 3MP cameras,” Maxon says. “The strong
resolution is across the entire range with the MegaDome all-in-one cameras have
resolutions from 1 to 5MP and the multi-sensor Arecont Vision panoramic cameras
have resolutions of 8MP across four sensors, at two megapixels for each sensor.
“We are currently
working on a series of projects in Australia in which we plan to deploy our 5MP
cameras in stadiums for crowd-control applications – the high resolution image
helps reduce the total camera count while providing the pixel density needed
for facial identification.
“Our 5 MP cameras
feature a maximum resolution of 2592×1944 and can deliver pristine image quality
with frame rates up to 9fps.”
According to
Maxon, Arecont Vision has successfully addressed what were previously labelled
as weaknesses in megapixel technology, such as bandwidth and storage issues,
low-light performance and high cost.
“H.264 compression
enables us to deliver megapixel video at similar bandwidth and storage
requirements as MJPEG VGA. Our day/night cameras provide very good low light
performance for night-time applications, and proper system design can take care
of any lighting issues with external white or infrared illumination,” he says.
“Related to
price, Arecont Vision’s products cost less than either IP VGA cameras or analog
cameras with encoders. One of our most important efforts is to match each
customer’s application and expectation of image quality to the right megapixel
camera.
“While our 5MP
camera is an excellent product in many projects, it is not necessarily ideal
for all scenarios. If the customer expects full motion, for instance, we would
steer them toward the 1.3MP or 2MP cameras. If the customer has a need to
capture content in low light, the 1.3MP, 2MP or 3MP cameras would be better
options, and our dual sensor day/night camera would be even better,” Maxon
explains.
“Being attentive
to customer requirements and setting expectations are important steps in
recommending the right megapixel camera.”
Over at Sony,
Tony Lagan says megapixel technology is considered as an image sensor with 1
million pixels or more.
“Most megapixel
cameras use an aspect ratio of 4:3, however it is important to note that HD
16:9 resolution of 1920 x 1080 is also Megapixel as it has approx 2.1 million
pixels,” he says.
“But the real
difference is the aspect ratio. 4:3 megapixel cameras are high resolution but
not high definition. To be classed high definition the camera must display an
aspect ratio of 16:9.”
Lagan agrees that
while the biggest advantage of Megapixel is resolution, the higher the
resolution generally, the higher quality the image.
“It makes
identifying people or objects so much easier,” he says. “Of course there’s an
obvious trade-off between resolution, frame rate and storage capacity with
megapixel IP. Because of the cost of storage and the limitations of networks to
handle the high bandwidth the best resolution would be between 1 to 3
megapixels. Obviously as compression algorithms improve and storage costs fall,
higher resolutions will be much more feasible.”
Megapixel and HD
each have advantages
Lagan think there
are very obvious advantages to both HD and megapixel he says where megapixel
takes the cake is thanks to its much higher resolutions.
“However, this is
not practical on a data network so really high resolution cameras are used in
isolation often with standard definition cameras,” he explains. “Meanwhile, HD
excels in its ability to effectively fill a 16:9 monitor without stretching or
cropping the image.
It is virtually
impossible to find a 4:3 monitor these days.”
For buyers, the
main things to take into account are to keep the limitations of megapixel in mind.
“By doing this
you can then design a system around the needs of each location. For example you
may want to use Megapixel or HD at entry points to capture faces but in another
location you may just need a camera to give an overview of the scene, therefore
SD would suffice,” Lagan says.
“Another
consideration is low light performance, as the major image sensor supplier to
the majority of the industry, what we have found is that generally the higher
the resolution, the less capable the low light performance.”
In terms of
standout features of Sony’s megapixel range, Lagan says there are three
features the company is really excited about.
“Firstly, there’s
our new View-DR wide dynamic range algorithm,” he says. “Where our competitors
take two images and combine them, the new Sony cameras take 4 images and
combine giving a much greater dynamic range resulting in images of incredible
quality in harsh back lit environments.
“The second is our
Visibility Enhancer which is a post-capture technology which dynamically
optimizes brightness by remapping, pixel by pixel, frame by frame. Near blacks
are expanded and near whites are compressed to provide better quality low light
images and also better performance in harsh backlighting.
“And the third is
our head end analytics, or DEPA,” Lagan explains. “This includes intelligent
motion detection, trip wires, loitering and people count.”
According Lagan,
Sony’s current benchmark megapixel camera delivers 3MP but he says as networks
and technologies improve then the company may look to expand this performance.
“What is
important to remember is that really high resolution image sensors are not
designed for video, and this effects reliability,” he explains.
“Don’t forget
that we have unique requirements in the security industry, that is that we
expect our cameras to operate 24/7 365 days a year. They are never turned off.
And we expect them to last for years.”
And Lagan says
while there are superior low light solutions available, system design can
resolve most these challenges.
“Low light is an
issue with megapixel but the Sony cameras utilise some unique features such as
“Light Funnel” and “Visibility Enhancer” to overcome this,” he says.
“SD cameras are
still king when it come to low light performance, however, by using IR lighting
high resolution cameras can get around this quite effectively.
“What is
important to remember is that really high resolution image sensors are not
designed for video, and this effects reliability. Don’t forget that we have
unique requirements in the security industry, that is that we expect our
cameras to operate 24/7 365 days a year. They are never turned off. And we
expect them to last for years”
Tony Lagan, Sony
