Modern trends in the development and use of infrared illumination in CCTV cameras.

It would be nice for a person to see at night as well as during the day. However, this idea can only compete in terms of utopianism with the possibility of growing wings. In the same way, creating a camera that gives the same colorful and bright picture at night as during the day remains a dream of developers and consumers. It would seem that technology does not stand still and such a seemingly not the biggest problem should be solved, but no, the quality of the night image cannot even compare with the picture that cameras give during the day. Is it possible to get an image from a camera in pitch darkness? Let's try to figure it out.
The first thing that comes to mind is to illuminate the observed object with artificial light, for example, gas-discharge lamps. But what if the object has linear dimensions of 100-200 m or an area of ​​about 2000 m2? It is easy to imagine that such an undertaking will not be cheap. Are there other ways? Yes: use infrared illumination devices. In this case, you can kill two birds with one stone: reduce potential costs, saving on lighting, electricity costs, and not create excess light, which is so feared recently in Europe by fighters against light pollution of the environment.
The essence of IR illumination is to illuminate the observed object in the dark with a source of infrared radiation. Infrared radiation is not visible to the human eye, but is perfectly recorded by television cameras. In this case, such wavelengths are selected so that the radiation spectrum is in the near infrared zone (wavelength of about 800-1000 nm: the so-called near IR range), where the sensitivity of the matrix is ​​still sufficient to obtain a monochrome image. The sensitivity of the camera to near infrared radiation makes it possible to use IR illumination (Fig. 1).

Fig. 1
The spectral sensitivity of CCD matrices characterizes the dependence of the matrix sensitivity on the wavelength of the received radiation. Compared to the human eye, the spectral sensitivity of most cameras is generally wider and extends into the infrared range up to wavelengths of about 1000 nm, if special filters are not used in the production of matrices. Thus, unlike the eye, the spectral sensitivity of which is limited to the visible range of radiation, the camera, due to its fundamentally different design, is also capable of registering the near range of IR radiation. Figure 2 shows the difference between the spectral sensitivity of the EXviewHAD matrix and the usual «non-EXview». It is clearly seen that the sensitivity of the EXviewHAD matrix in the IR region starting from 740 nm is twice as high as that of the usual one. In other words, in the dark, when illuminated by IR light of the same intensity, the HAD matrix image will be much more detailed and contain less noise, and, conversely, the same image quality will be achieved with half the illumination of the object in front of the HAD matrix.

Fig. 2. The spectral sensitivity of color cameras is closer to the human eye. This is explained by the design features of the color matrix and the use of an infrared filter that cuts off infrared radiation. The area of ​​each light-sensitive cell of a color camera is several times smaller than that of a black-and-white camera, since each such cell is a triad, which consists of three light-sensitive subcells corresponding to the three color components of the image: blue, green and red. Accordingly, with the same dimensions of the black-and-white and color matrix, the latter will have a smaller subcell area. In addition, each subcell of the color matrix has its own light filter (green, red or blue), which weakens the overall light flux directed to the matrix. As a result, color TV cameras are characterized by lower sensitivity compared to black-and-white cameras and no sensitivity in the infrared part of the spectrum.
For color cameras used in television surveillance systems and operating in the visible part of the spectrum, special measures are taken (using infrared filters) to limit the spectral characteristics and bring them to the visible part of the spectrum. If there is no filter in front of the camera lens that blocks the IR spectrum, the deterioration in image quality will be immediately noticeable: the contrast decreases, noise and color distortion appear. In day-night cameras, a solution in the form of a mechanical IR filter is also used: during the day, when there is sufficient illumination, it is installed and the camera shoots in the visible spectrum, and at night, when there is not enough illumination, the camera switches to black and white mode, the filter is removed and the IR illumination is turned on.
Structurally, an infrared filter is a small plate of glass or any translucent polymer with a layer applied to it that absorbs IR radiation. Ultimately, the filter can be simply a thin film of a material that reflects infrared radiation, applied to the lenses of the camera optics or to the CCD matrix itself.
Figure 3 shows the dependence of the transmittance on the wavelength of light of a typical IR filter. It is evident that the maximum transmittance (about 85-90%) falls on the visible spectrum, and the minimum — on the near infrared and ultraviolet ranges, where the IR spectrum is weakened by 10-15 times.

Fig. 3

Main properties and parameters of infrared illuminators in video surveillance systems
Let us consider the main characteristics of LEDs that determine the parameters of infrared emitters they are a part of. Devices with infrared radiation have the same characteristics as other light sources. One of the main parameters of LED that interests us is the wavelength of the emitted IR light. Since CCD matrices usually have a decrease in sensitivity with an increase in wavelength to the infrared range, LEDs are usually selected whose main radiating ability falls on a wavelength of 850 nm. These LEDs can be seen to glow reddish in the dark, because their spectral characteristic partially falls within the visible spectrum. LEDs with a maximum spectral characteristic at 930–950 nm have completely invisible radiation. If there is no need to organize hidden IR illumination when organizing video surveillance, then, of course, you should not try to install illuminators in the 930–950 nm range, since the sensitivity of the matrices in this area is lower than in the 830–850 nm range.
The axial luminous intensity measured in candelas largely determines another parameter – the illumination range. The illumination range of an entire device consisting of multiple LEDs can be increased by using more powerful LEDs or by increasing their number. However, increasing the number of LEDs, as well as increasing the current passing through each LED of the IR illumination, leads to an increase in the detection range only up to a certain point – reaching the so-called saturation region. The illumination range (Fig. 4) can be estimated, for example, as the distance from the emitter to the vertical plane, determined from the condition of minimum illumination equal to 2 lux, the point of intersection of the plane and the emitter axis (for a camera with a sensitivity of at least 0.1 lux).

Fig. 4

Determining the illumination range of an infrared emitter.
An infrared illuminator has such a property as the distribution of the luminous flux in space, which can be represented as a diagram (Fig. 5) of the spatial distribution of luminous intensity (directional diagram). This diagram shows the share of the emitted energy in the selected direction from the total intensity. In other words, how effectively the emitter concentrates the luminous energy in the desired direction and scatters it as little as possible in other directions. The radiation angle of the light source is determined by the directional diagram as the angle formed by the rays emerging from the point source, passing through the intersection points of the directional diagram and the line determining the level of half the relative axial luminous intensity.

Fig. 5

An accurate diagram of the spatial distribution of luminous intensity can be constructed using a special device — a goniometer. A rough directional diagram can be obtained using a photo sensor, manually rotating it in a horizontal plane along the radius of a circle in the center of which is the LED, and recording the output level and the angle of rotation of the sensor. However, one cannot count on high accuracy of this method.
If we consider not the entire emitter as a whole, but only one individual diode, then a typical directional diagram for it will look approximately as follows (Fig. 6).

Fig. 6

Some LED manufacturers provide approximately this or a similar picture as a graphical representation of the LED directional pattern. Deviations in lens geometry, errors introduced during production, and aging of the housing material over time can significantly affect the optical properties of LEDs (Fig. 7). It is worth considering that the unevenness of the IR emitter directional pattern strongly depends on the emitter design, its production quality, and the LED radiation angle.

Fig. 7

A separate topic for discussion is the durability of LEDs or their mean time between failures. It is desirable, of course, that this time is not less than the estimated service life of the camera. Service life is an important (and painful) operational parameter of semiconductor light sources. Here, two criteria can be distinguished: full (until the device completely fails) and useful (until the luminous flux falls below a certain limit) service life. When designing IR illuminators, one must not forget about their further operation, in particular, about the possibility of replacing the IR backlight unit in outdoor cameras.

The main reasons for failure and reduction of the service life of IR diodes
Degradation of the active region
Light emission in a LED occurs as a result of recombination of injected carriers in the active region. The emergence and growth of defects in the semiconductor crystal lattice leads to its degradation. Physical processes occurring in the semiconductor (high density of the injected current and the associated heating of the semiconductor) inevitably accelerate the development of the defect. But since defects in the crystal lattice are present in all semiconductor devices, all LEDs are subject to the development of degradation of the active region. Increased supply voltage significantly contributes to this.
Thermal degradation
The amount of heat generated by IR LEDs requires that they be mounted on a radiator. Overheating of the semiconductor leads to an increase in the concentration of minor charge carriers (electrons in the p-region and holes in the n-region), which form a reverse current that is highly temperature-dependent. Uncontrolled temperature growth can cause failure of the LED semiconductor due to thermal breakdown. Long-term operation at elevated temperatures leads to thermal destruction of the semiconductor's crystalline structure due to an avalanche-like increase in the number of atomic migrations from one region to another.
Electrostatic discharge
Semiconductors are very sensitive to electrostatic discharge. ESD damage can manifest itself as a sudden failure or internal damage, leading to rapid failure during subsequent operation. According to existing regulations, the sensitivity of LEDs to electrostatic discharge should be at least 100 V. Breakdown due to static discharge can be a significant problem for LEDs used in IR illuminators, since the power cable can collect static electricity during a thunderstorm, and no arresters are usually provided by manufacturers for IR illumination.
The following conclusions can be drawn from the above. To increase the service life and prevent failure of infrared backlight LEDs (and any LEDs in general), it is necessary to observe the thermal conditions of their operation (use thermal insulation of IR backlight units, stabilize the nominal supply voltage) and provide protection against surges caused by static electricity.
Most LEDs supplied by the industry today degrade to varying degrees within a few years, despite the fact that manufacturers usually guarantee the service life of their LEDs at about 100,000 hours or about 11 years of continuous operation. However, these are only theoretical calculations, which assume that the diode will operate in greenhouse conditions. In reality, everything is much more complicated. An analysis of numerous failures of IR illumination devices allowed us to draw the following conclusions. Perhaps the most destructive factors are poor assembly or production quality, temperature and instability of the supply voltage. Assembly quality determines the durability of operation by 40 percent, extreme and elevated temperatures reduce the service life of the LED by another 30%, instability of the supply voltage takes away 20% of the estimated service life. Factors such as defects in the semiconductor crystal lattice, electrostatic discharges, material fatigue and all kinds of mechanical stress in the device have a complex effect on the service life of the LED, and it is impossible to take them into account, but all together they provide the remaining 10% of reliability.

The feasibility of using day-night cameras together with built-in IR illumination
Night video surveillance using separate LED IR spotlights in general and built-in illumination (again LED) in All-in-One cameras is probably the most acceptable in terms of image quality/price, when compared to cameras with accumulation and artificial lighting. Again, even if a camera with accumulation produces a static image of similar quality, its use in places where it is necessary to shoot moving objects is completely pointless due to its ability to turn moving objects into translucent ghosts.
If we consider surveillance using IR illumination as an alternative to surveillance using artificial lighting, then from an economic point of view the first method is unrivaled: the cost of lighting equipment and energy consumption is several times (if not an order of magnitude) higher than the use of cameras with IR illumination. Simple calculation: on a perimeter length of 100 m, you can install, for example, 3 lanterns of 500 W or 4-5 IR illuminators of 10-15 W — the savings are obvious.
Hidden lighting does not draw attention to the hidden video camera, which allows for more successful counteraction to, for example, deliberate theft of equipment.
It is also necessary to highlight two conditions for using IR illumination. Firstly, the unmasking glow of the radiation sources themselves is acceptable. In this case, it is possible to use emitters with a wavelength of 920, 880 and 850 nm. Almost all LEDs emitting in the 840–880 nm range have a sufficiently intense visible component of light red color, especially at maximum currents. Secondly, the emitter itself must be absolutely invisible even when directly visually observed from a close distance. Emitters with a wavelength of 930–950 nm are used for this. But it should be taken into account that the sensitivity of the CCD matrix to wavelengths of 840–880 nm is higher than to wavelengths of 930–950 nm, therefore, all other things being equal, the use of illuminators in the 840–880 nm range will be preferable.As for the comparison of external illumination and IR illuminators built into the camera housing, not everything is clear. An infrared illuminator can provide a more powerful and uniform radiation flow, but at the same time it consumes more power, requires additional installation, and is not compact. Built-in illumination is compact, does not require additional installation and is quite economical. But under certain conditions, the built-in illumination source can create glare and flare of the camera lens, and the spatial distribution of light often leaves much to be desired. Another interesting fact: the coordination of the illumination and camera. In All-in-One cameras, as a rule, the IR block and mechanical light filter work in concert: the illumination is turned on and the filter is removed simultaneously. While a separate IR illuminator with a built-in photo sensor can turn on later than the filter in the camera is removed, then the camera «goes blind» in the dark for some time. Let's consider various typical solutions available on the market today as examples of the implementation of All-in-One cameras.
The simplest type of camera with IR illumination can only be called a toy — 6 LEDs have low radiated power and an uncoordinated directional pattern, so such illumination is effective at a distance of more than 2-3 m, and only for the central part of the frame. Cameras in casings are now quite common. Behind the glass shared with the camera there is an illuminator made of several dozen IR diodes, this glass is the main disadvantage of such a design: part of the IR radiation will inevitably get into the lens due to all sorts of reflections inside the casing, and the better the parts of the camera and the light-protective hood are fitted, the less the illumination will be.
Even greater problems with illumination may arise during operation with outdoor sealed dome TV cameras with built-in IR illumination. It is difficult to achieve tight contact between the lens and the hemispherical dome with such cameras, and the dome itself is a monolithic hemisphere, in which the parts of the common glass covering the lens and the illumination unit are not separated in any way to prevent illumination of the camera. After some time, dust and dirt settled on the dome will become an additional source of illumination, reflecting infrared light back into the lens, and after a couple of months the consumer will be perplexed by the glow of dirty glass instead of a picture.
Another negative factor of the built-in illuminator is the interfering backscattering of the environment at high density: snow or rain. The simplest method of minimizing it is the use of distributed illumination — separate illuminators that create a uniform light field in the camera's field of view.
In general, it can be concluded that, despite the convenience of using built-in IR illumination for TV cameras, it can only be recommended for solving simple and not very important tasks. In cases where increased requirements are imposed on video surveillance at night or in low light conditions, the most justified is the use of separate illuminators with directional patterns matched to the field of view of the TV camera. In the same way, to create a relatively uniform light flux and reduce backscattering of the environment, it is advisable to use external illuminators located next to (usually symmetrically on the sides) the camera.
Recently, manufacturers of All-in-One outdoor cameras, in the conditions of competitive struggle for the market, have been improving the quality of camera shooting in night mode using infrared lighting. One of these areas is the use of adaptive illumination technology, designed to level out the illumination of objects located close to the camera. Its essence is simple: the light intensity of the built-in IR illuminator is adjusted according to the illumination level of the observed object, i.e. a kind of optical feedback is formed between the matrix and the illumination unit.

Let's sum it up. The best way to implement video surveillance at night and in low light conditions is to use external IR spotlights and illuminators. All-in-One systems using simple IR illuminators are probably the best choice, considering the price/quality ratio. In addition, the industry does not stand still and offers new viable solutions, such as, for example, adaptive illumination. And this allows us to hope that All-in-One systems, despite their inherent shortcomings, will be very popular in the foreseeable future.

References Chura N. I. Some aspects of using IR illumination in video surveillance //Special equipment. 2002. No. 3. pp. 35–39. Chura N. I. Some aspects of using IR illumination in video surveillance //Special equipment. 2003. No. 5. P. 35–38. Nikiforov S. G. Problems, Theory, and Reality of LEDs for Modern High-Quality Information Display Systems //Components and Technologies. 2005. No. 5. P. 48–57. Nikiforov S. G. Why Don’t LEDs Always Work the Way Their Manufacturers Want Them? //Components and Technologies. 2005. No. 7. P. 16–24.