Wider world: ultra wide angle surveillance

15 Dec 09

By Dave Tynan

With a variety of wide angle cameras on the market, it is important to understand the differences in capabilities, says Dave Tynan of Avigilon...

Since their introduction several years ago, megapixel and HD surveillance cameras have offered the industry a powerful tool for covering large areas in high resolution. Now in mainstream use, megapixel IP surveillance cameras reduce system costs by replacing multiple cameras and improving investigations by capturing better, and more useful, evidence.

Developed specifically to cover ultra-wide scenes with fewer cameras, wide angle surveillance cameras are increasingly available.

But with a variety of these cameras on the market, it is important to understand the differences in capabilities and results in order to design a high definition surveillance system built to meet your specific needs.

Ultra wide angle advantage

The need for ultra wide angle surveillance cameras is evident when we consider the surveillance requirements of a 10m by 10m building lobby. Using a traditional camera environment would necessitate the installation of a camera in every corner for complete coverage, particularly if there are multiple entrances or areas of activity along each wall.

In a situation like this, purpose-built megapixel surveillance solutions targeted at 180 or 360 degree imaging can provide the coverage required to monitor the entire area, with fewer cameras.

If central mounting is possible, a single 360 degree camera can give you complete coverage, but if it isn’t practical to mount a camera in the centre of the room, two 180 degree cameras on opposing walls can deliver the same coverage.

In both cases, the number of cameras required is reduced along with the associated installation and service costs.

Covering 180 and 360 deg

There are currently two options available on the market for gaining 180 and 360 degree coverage from a single camera: a specialized lens on a single sensor or a camera that integrates multiple sensors with conventional lenses aligned to provide ultra-wide angle coverage.

Single camera solutions for ultra wide angle coverage use a specialized fisheye lens. To enable ultra wide coverage, fisheye lenses use different optical design techniques compared to conventional lenses, which can often lead to the distortion of the captured image when viewing beyond a 90 degree horizontal field of view.

In some cases, a ‘barrel’ distortion, where a circular image is created and straight lines within the captured image appear curved, can occur. This optical distortion can sometimes be corrected in software, but is typically a very processor-intensive task that requires proprietary algorithms.

Another limitation of the fisheye lens is the way it projects the captured image onto the image sensor.

As a consequence of its design, it will create either an oval or circular shaped image, which can be problematic because all image sensors used in surveillance cameras are square or rectangular.

As a result, a 180 or 360 degree fisheye lens is never able to use all the pixels available in the sensor, as you can see in the example image in figure 1

Recent adoption

Fisheye lenses have been used in photography since the early 1900s, but their broader adoption in the surveillance industry hasn’t occurred until fairly recently, in part because of the introduction of IP HD and megapixel surveillance cameras during the last decade.

Covering an ultra-wide 180 or 360 degree field of view with conventional analogue technology was ineffective because the resulting images had no detail.

The second option that can be used to achieve 180 or 360 degree coverage indoors is a compact, multi-sensor camera that uses conventional lenses (figure 2).

Typically, these cameras will use four 2MP sensors with compact, pre-installed conventional lenses.

In a four sensor, 180 degree multi-sensor camera, each lens covers a 45 degree field of view.

With a four sensor 360 degree camera, each lens covers a 90 degree field of view.

The advantages of a multi sensor approach are clear: the images are captured without distortion and deliver an easily measured and quantifiable amount of pixels-on-target consistently throughout the scene.

Dramatic difference

Compact multi-sensor cameras typically have a dome form factor and use one network cable for power and to transmit video for the images from all four sensors, making installation quick and convenient.

And because conventional lenses are used, every pixel on each sensor is utilized to capture the image, maximizing overall image quality without wasting any pixels. The difference in detail, security, and coverage for the two approaches can be dramatic.

Look again at the examples of a 10m by 10m building lobby with a single 360 degree camera mounted in the centre.

If you use a circular fisheye lens and a 3MP camera with a resolution of 2048 x 1536, you capture the room with a distorted circular image and a diameter of 1536 pixels. That circular image has a circumference of 4825 pixels covering the 40 meter perimeter of the room, or roughly 120 pixels per meter at the edge of the room.

A multi-sensor 360 camera using four 2MP sensors has 6400 pixels of total horizontal resolution across the four streams, resulting in 160 pixels per metre of detail around the perimeter of the room.

Faces at the perimeter

In this example, only the multi-sensor solution provides sufficient detail to identify faces around the perimeter of the room. The other advantage of the multi-sensor solution is that the resolution in the balance of the image can be easily and reliably calculated. This is done using a simple formula dividing the width of image sensor by the scene covered.

The circular projection of the fisheye lens means this math can only be done easily at the perimeter of the image. The distortion of the circular fisheye image means that you have the least amount of detail at the perimeter and the most in the centre of the room.

Concentrating image detail in the centre of the image in most configurations is typically wasted because a single sensor fisheye camera will only capture the tops of heads directly beneath the camera.

Pixels on target

The introduction of megapixel surveillance cameras has improved these limitations, but there are still important concerns to be aware of when coupling ultra-wide angle lenses with megapixel surveillance cameras.

Any modern surveillance system design must now address the concept of pixels-on-target – the ability to quantify the detail of the image being captured during design and verify it during installation.

Pixels-on-target, a simple concept pioneered by Avigilon and other vendors in the HD and megapixel surveillance space, ensures reliable system design by providing a verifiable metric on image detail.

Unfortunately, pixels-on-target can only be easily calculated when conventional lenses are used, because ultra wide angle lenses distort the scene when captured. As a result, you can no longer guarantee the amount of pixels-on-target at any point in the image.

Recording and management

Finally, effective recording and management is always an important consideration for any video surveillance technology.

Effective coverage of ultra wide angle scenes is a great way to leverage the resolution of multi megapixel IP surveillance cameras to reduce system costs. However, multi megapixel surveillance cameras pose unique challenges with bandwidth, storage, and management that needs to be properly considered when selecting a HD and megapixel surveillance system.

Purpose-built network video management software (NVMS) platforms are now available that are tailored for HD and megapixel surveillance and can be used to ensure all the evidence captured by a 180 or 360 camera is stored and managed effectively.

Knowing what options are available and understanding the advantages and disadvantages of each is the first step towards designing a high-performance, cost-effective high definition surveillance system.

With a clear understanding of the approaches taken, integrators and end-users are in a better position to build a system that offers complete coverage and unmistakable evidence for improved investigations.

Credits :

Dave Tynan is vice president of global sales and marketing at Avigilon