Under cover of darkness

Visual information is invaluable in a wide range of applications, and the positive elements of using CCTV – the ability to remotely observe buildings, grounds, people and public areas – is widely understood. But there is one element that is necessary for visual information to be gathered, without which the information is next to useless. That element is light – and a lack of the required level of light can be disastrous for even the most advanced CCTV system.

The reflection and refraction of light are basics of physics. More importantly, without these processes, we would be unable to define any objects or activity visually. Without being artificially created, there are very few places on earth that have no light whatsoever, which is why we can often see in very low light levels, once we have become accustomed to the darkness. This is because of the very advanced and complex processing that goes on between the eye and the brain. Unfortunately, CCTV devices do not share that phenomenal processing power!

CCTV systems ultimately face restrictions. As we’ve mentioned, they lack the huge intelligent processing capability of the human brain, and the advanced mechanics of the human eye. Surveillance equipment is also built to a cost, meaning that even if such capabilities could be replicated, the cost could turn out to be much higher than the CCTV industry would accept.

However, by using what technology we have in the industry, steps can – and have – been taken to ensure that usable video images are obtainable in low light environments.

Considerations

Before an effective choice of camera for a 24-hour CCTV system can be made, there are many factors that need to be considered.

Firstly, the objectives of the system should be identified. For example, what quality recording is required? In many systems, the recording system will denigrate the image quality available from the camera, because archiving is required for 31 days, so resolution and frame rate levels are compromised.

It is therefore essential to get the best possible image to start with. Technology such as SSNR (Samsung Super Noise Reduction) will help in this regard by filtering the image to provide cleaner images with smaller file sizes, to optimise capacity of DVRs. The camera that can give perfect, noise-free images in pitch darkness is not yet available for day to day security systems, so light sources that are available at night, if any, in addition to moonlight and starlight, will dictate the type of camera used.

Common light sources such as low-pressure sodium (street lights), tungsten-halogen (flood lights) and infrared lamps, can all be used to balance light pollution with available camera technology, to meet the system objectives.

Major objectives

There are two main objectives when considering creating a CCTV system for low light environments. The first is to try and channel as much light as possible onto the CCTV camera’s CCD chip. The second is to minimise the negative effects created if a sufficient level of light does not actually reach the CCD.

The majority of modern CCTV cameras use a CCD to produce the images that are viewed on the monitor screen or recorded on a DVR. The technology used by CCDs has been developed over many years, and manufacturers of chips are still looking for new ways to make the chipsets deliver better performance, especially where lighting levels are not consistent or guaranteed.

A CCD, or charge coupled device, is a very small light-sensitive silicon chip that has an area on the surface that absorbs light. This area is split into a number of cells, known as pixels, which means that a CCD chip is really an array of light sensors. These sensors are coupled to allow the controlled movement of charges, which build

up in each pixel when light is channelled into the camera, focused onto the CCD via the lens.

The light that falls on the pixels causes them to react proportionally to the luminescence detected, and an optical image of the viewed scene is created. Obviously, if light levels are poor, the amount of light focused onto the CCD will be low, and the subsequent video signal will be weak and susceptible to lack of fidelity and contrast, reduced brightness, and noise.

A CCD-generated image will be made up by between 250,000 and 500,000 pixels. In terms of sensitivity to light and picture resolution, the CCD is much better than the technologies it replaced.

By way of an example of the performance CCDs can achieve, a monochrome camera such as Samsung Techwin’s BW-4302 has a resolution of more than 570 TV lines, and is sensitive enough to deliver images of an acceptable quality at light levels as low as 0.05 lux.

A day/night camera – the most commonly used option for those who need constant video surveillance, even when the level of illumination falls greatly, as at night – such as Samsung Techwin’s SDN-520, gives 480 TV lines of resolution in colour mode and 520 TV lines of resolution in monochrome mode. It can deliver images in light levels down to as low as 0.01 lux.

This is surpassed by the SHC-721A, which gives 500 TV lines of resolution in colour mode, and 550 TV lines of resolution in monochrome mode. It can also deliver images in light levels down to as low as 0.01 lux.

The latter two units also feature noise reduction, eliminating one of the causes of poor video quality in dark environments.

To put these figures into context, the 1.0 to 10.0 lux range is equivalent to the light emitted at dusk and dawn, while quarter moon light is just 0.1 lux. Therefore, a camera with a sensitivity as low as 0.01 lux will still deliver video with a strength of 1V peak-to-peak in very poor lighting conditions.

Low-light cameras will typically incorporate a feature called frame integration, or ‘Sense-Up’, which enables the camera to build up layers of images like the slow shuttering on a still camera. This technology enables images in conditions where no ancillary lighting can be used, down to 0.01 lux (starlight).

As with still cameras, there is a downside to this technology. It is great for seeing stationary objects in near darkness, but any movement will ‘ghost’ and leave a trail behind it. Most sense-up cameras will have an automatic setting that will only come into play when the light level falls below that required for normal operation and a maximum setting can be applied to restrict the degree of ghosting to that acceptable to the individual user.

Types of camera

It is perhaps worth clarifying the actual function of some popular names for different low light camera types, in order to specify the right camera type with the right type of ancillary lighting.

Colour/mono cameras switch between colour mode and monochrome mode, and have been used for quite some time to provide 24-hour surveillance. They have often given poor colour rendition, because the sensitivity of the CCD sensor to infrared affected the colour image, even though infrared response is reduced.

They are, however, very well priced, and latest generation cameras have advanced image and noise reduction processing to give good daytime colour rendition, clear, low-noise colour, and excellent low-noise monochrome images at low light (0.3 lux). They are best used in areas with some ambient lighting of most types, although IR lamps should have a cut-filter of 715nM or below.

Day/night cameras are similar to colour/mono, but use a switchable IR cut-filter in the camera to give the best possible colour images in the day, and maximum IR sensitivity at night. These cameras offer the best of both worlds without compromising performance, but at a slightly higher cost. They can be used with any common light source, and are ideal for use with IR lamps up to 900nM.

An important element when selecting the appropriate camera is its spectral range and how it matches the wavelength of the light source.

If there is existing IR lighting on site, then it is essential to ascertain the light wavelength of the lamps and check that it is within the spectral range of the chosen camera.

Maximising performance

Once you have a camera that is suitable for the job, the next thing to consider is how to maximise its performance.

The selection of a lens is important, especially if a day/night camera is being used. For example, if a standard lens is used, back focus might be compromised. Visible light and infrared light have different frequencies, and this causes a focus shift as a camera switches from colour to monochrome mode. This means the focus cannot be correct in all conditions, so it is always worth considering the use of infrared corrected lenses.

Also, consider the use of aspherical lenses, which are designed to let more light be focused onto the CCD. Using low shutter speeds lets more light in, but targets moving at speed could be blurred.

Increasingly, cameras also make use of technologies such as frame integration. However, when you look at images that are captured in very low light situations, the thing that usually makes the picture unsuitable is the amount of noise that it carries. Noise occurs when AGC (Automatic Gain Control) attempts to compensate when the video signal is weak, but it is usually the graininess and snow-type effects than render an image unusable before the actual darkness does.

If the noise could be removed, the image would still be usable. This is the thought behind the development of SSNR, which makes images cleaner and sharper, eliminating the high levels of noise without creating ghosting or motion blur.

Independent tests have shown that colour images with SSNR applied stay sharp and warm, even after colour information is lost due to lack of light. This shows that SSNR reduces noise in colour images that are well below the usual standard signal strength of 1V peak-to-peak, without the pictures becoming unusable for CCTV purposes.

The technology is even more impressive when the camera is in monochrome mode. If a CCTV system is surveying an area where crisp, clean and detailed images are required in all types of illumination, cameras with SSNR can solve many problems that standard cameras will struggle to overcome.

Because the processing increases the signal to noise ratio and considers both random and predictable sources of noise, this even means that streaking from vehicle lights in motion can be eliminated. SSNR, coupled with other, more traditional low light features and functions such as Sense-Up and AGC, can really enhance images that otherwise would leave a lot to be desired.

Great expectations

Customers’ expectations in respect of the performance of CCTV continue to grow – regardless of the application. Having to use CCTV in low-light environments is therefore a reality, and increasingly something your customers will demand. There isn’t a single answer to the difficulties presented by low light environments, but creating a system for such circumstances is not a hard task.

Start with a good quality camera that can deliver solid low light performance in the prevailing lighting conditions. A combination of high resolution and low sensitivity is vital. Also ensure that it has the required features and functions, such as AGC and Sense-Up.

Ensure that the lens used matches the performance of the camera, getting as much light as possible onto the CCD. And remember that noise will often be detrimental to the final obtained image. With noise reduction techniques such as SSNR, you can create CCTV systems that give excellent recorded results, even in the poorest level of illumination, and optimise the available archive days from your DVR.