Thermal imagers: it's not that simple.

Thermal imagers: it's not that simple.

Thermal imaging is becoming one of the hottest topics in various industry publications.

We read several articles on thermal imaging technology and were surprised: most authors pay significant attention to the intricacies of the design of microbolometers, someone uses drawings to prove that indium antimonide is more suitable for infrared photodetectors than silicon.

One ​​of the most widely circulated statements (or misconceptions) is that thermal imagers should be used everywhere, from deer hunting to inspecting gas pipelines for leaks.

This is all very well, but what should an engineer do who needs to create a thermal imaging system? Go to an «expert»?

Are you sure that he will offer a solution to your problem, and not try to solve the problem of fulfilling his sales plan?

You can turn to search engines, but it is easy to get lost in the huge amount of information that has accumulated over 60 years since the use of infrared cameras began.

Therefore, in this article we decided to touch on issues that will be useful primarily to engineers — indeed, the correct use of thermal imagers can solve many problems that the world poses to us.

It is difficult to explain the principles of quantum physics used in thermal imaging in a magazine article. And is there a need for this?

After all, most readers will not be developing thermal imaging cameras.

We decided to give clear and practical advice that may be in demand primarily when designing complex security systems.

We will consider three main topics:

1. Propagation of IR radiation.
2. Cooled thermal imagers – a necessity or “budget spending”?
3. Why are special camera characteristics used in thermal imaging?

Let us reassure the reader in advance: the laws of thermal imaging are very similar to the laws of television.

Only in the first case we work to a greater extent with the bodies’ own radiation, and in the second – with the reflected radiation of the Sun or other light sources.

In the first case we are dealing with active optical location, in the second – with passive.

The main task of CCTV is to assess the situation, while thermal imaging is more often assigned detection functions. Therefore, thermal imaging is in many ways similar to the CCTV we are used to, but still different from it.

QUESTION ONE. WHICH WINDOW SHOULD I LOOK IN?

For some reason, almost all modern authors omit this question. But this is as wrong as immediately starting to study Einstein's theory of relativity, turning a blind eye to Newton's physics! It is not for nothing that the classics of thermal imaging [1, 2] begin their books by examining this topic. Everyone knows that visible light, which our eyes and television cameras see, occupies only a small part of the electromagnetic spectrum. IR radiation differs from visible light in its longer wavelength, but the general principles of electromagnetic wave propagation are very similar. At the same time, there are a number of features that should be taken into account when creating a thermal imaging system. The first is the presence of «windows» of atmospheric transparency. But they say that a thermal imager is remarkable for the fact that it can see through snow, smoke, etc. — is that really not true? In general, yes. But there are details.

Thermal radiation is weakened as it passes through the atmosphere due to absorption by gas molecules, aerosols, precipitation, as well as smoke, fog, smog, etc. The following substances (listed in order of importance) absorb IR radiation in broad bands centered at the indicated wavelengths:

1. water (2.7; 3.2; 6.3 μm);
2. carbon dioxide (2.7; 4.3; 15 μm);
3. ozone (4.8; 9.6; 14.2 μm);
4. nitrous oxide (4.7; 7.8 μm);
5. carbon monoxide (4.8 μm);
6. methane (3.2; 7.8 μm).

Apart from attenuation in dense dispersed media, molecular absorption is the main cause of radiation attenuation, with radiation being most strongly absorbed by water vapor, carbon dioxide, and ozone. In the lower layers of the atmosphere, absorption by nitrous oxide and carbon monoxide can usually be neglected. Thus, taking into account the above, the positions of two transparency windows can be determined: 3.5-5 μm and 8-14 μm [3].

In practice, the presence of transparency «windows» means that all thermal imagers must operate in these ranges. The short-wave (3-5 µm) range is more typical for cooled thermal imagers, the long-wave (8-14 µm) range is more typical for uncooled (Fig. 1). Why? It's quite simple. Our world is designed in such a way that different devices are needed for high-quality detection of IR radiation in different parts of the spectrum. In the short-wave range, receivers with a photoelectric effect are used — the quantum energy is sufficient for electrons to move into the conduction zone under the influence of IR radiation. In the long-wave range, bolometers are used much more often, since it is easier to detect radiation in this part of the spectrum using the thermistor effect. A reasonable person tends to spend a minimum of effort to solve problems, so the equipment should be chosen consciously. But you can say, what if a cooled thermal imager is so much better that it is not a sin to pay for it? Let's figure out what is better and why.

QUESTION TWO. COOLED OR UNCOOLED?

We would like to warn the reader right away: we are not afraid of the price of cooled thermal imagers! We do not understand why the vast majority of articles either omit this type of equipment or say the following about it: “Guys, they are so expensive that you will definitely not buy them from us.” We believe that this is simply unfair — as if you come to a store and the seller tells you: “Step away from this fresh fillet — you can’t afford it!” Our goal is to tell engineers about the entire range of thermal imaging equipment, regardless of its price.

The price of cooled thermal imagers is quite comparable to the price of uncooled ones, if you add to the latter, for example, the cost of a new SUV — but this is not the pricing factor. The reason for the high cost of the device is the high cost of semiconductor matrices and cooling devices to ultra-low temperatures. But sometimes only a cooled thermal imager can solve the task.

In order not to overload the reader with unnecessary information, we decided to focus on the main advantages of both types of thermal imagers, without ignoring their disadvantages. After all, often knowledge of the weak point of a particular equipment will help to avoid further difficulties. The main advantages of cooled thermal imagers are:

• Better resolution – they operate in a shorter wavelength range compared to bolometric thermal imagers. According to the Rayleigh criterion, the resolution is determined by the ratio R=D/1.22l, where D is the lens diameter and l is the wavelength. The angular diffraction limit (this term refers to the minimum angular size of a monochromatic source) of a cooled thermal imager is wцl/D, where l is the wavelength and D is the lens diameter, is about 0.08 mrad (0.004 degrees). For uncooled thermal imagers, this parameter is 3-4 times lower.
• Cooled thermal imagers have greater contrast sensitivity — a cooled thermal imager can distinguish differences of 20 mK with an aperture of 5, while an uncooled bolometric one can distinguish about 50 mK, provided that the aperture is equal to one. This is a consequence of the different physics of the photoelectric and thermistor effects.
• The combination of the first two factors gives the third advantage — a much greater detection range. A detection range of 10 km is far from the limit for a cooled device.

Speaking of the advantages, it is worth mentioning the disadvantages of cooled systems:

• High power consumption due to the presence of cooling devices compared to uncooled devices.
• Relatively long cooling time — several minutes may pass between turning on the thermal imager and obtaining an image.
• Limited service life due to the mean time between failures of the cooling element — usually several thousand hours of continuous operation.

Let's now consider uncooled thermal imagers. The main advantages of uncooled thermal imagers are:

• The operating range is better suited for observation in conditions of smoke, fog, smog — in the range of 8-14 microns, IR radiation is not absorbed by either water vapor or carbon dioxide (the transparency window is “more transparent” than in the range of 3-5 microns).
• Relatively small size and weight.
• Uncooled thermal imagers work immediately after switching on. They are also characterized by lower power consumption.
• Very long mean time between failures.

The main disadvantage of microbolometric thermal detectors is the requirement to use high-aperture optics — for the thermoresistive effect to occur, it is necessary to emit and transmit a large amount of energy to the bolometer. Therefore, to achieve the required signal-to-noise ratio at the output of the photodetector, optics with a large entrance pupil diameter are required. But humanity has not yet created lenses with a relative aperture significantly smaller than 1, and it is unlikely that it will learn to make them in the foreseeable future. It is the physical boundaries of what can be implemented in optics that have always been a limiting factor in the applications of infrared television — from fire detection to ballistic missile launch detection.

Another nuance lies in the emissivity of bodies heated to different temperatures. Wien's law gives an idea of ​​​​where the maximum emissivity of a particular body is located. For example, we can calculate the following important maximums for us:

1. Human – 9.36 microns.
2. A car or boat with an internal combustion engine – 8.45 microns.
3. Forest fire – about 3 microns (depending on the nature of the fire).

But it should also be remembered that the maxima of the thermal radiation of the object are located at these wavelengths. Both a person and a car have a noticeable luminosity in the short-wave range, so thermal imagers operating in the 3-5 micron range can be used to observe them at long distances. Bolometric thermal imagers cannot solve this problem due to the above-mentioned physical limitations.

QUESTION THREE. WHY ARE NEW EQUIPMENT PARAMETERS NEEDED?

We are all accustomed to the fact that most television cameras are characterized by the signal-to-noise ratio. To obtain a high-quality image, the signal-to-noise ratio at the output is approximately 50 dB. Using simple formulas from a textbook on radio engineering, we can even say that this means: the signal power is greater than the noise power by approximately one hundred thousand times. It seems that everything is clear, the same parameter can be used to evaluate thermal imagers. However, manufacturers of infrared equipment indicate not only the signal-to-noise ratio that we are accustomed to, but also introduce some other parameter NEP. Why?

The noise equivalent power (NEP) is a measure of the sensitivity of an optical receiver. It is defined as the signal power that produces a unity signal-to-noise ratio at the output of the optical receiver for a given operating wavelength and effective bandwidth. Naturally, it is desirable to have the lowest possible value of the equivalent noise power, since in this case the signal-to-noise ratio will be the highest. Using NEP instead of signal-to-noise ratio to describe a receiver is preferable because NEP remains constant under different conditions. For example, if we remove the receiver from the source by 500 and 1000 meters, the noise level can change either up or down (Fig. 2).

It is clear from the figure that in the first case we have one signal-to-noise ratio, in the second case – a completely different one, because the source signal is weakened, say, by a cloud of steam (indicated by a blue triangle), and the noise is amplified – a background source, for example, trees heated during the day, makes its contribution. One of the main quality criteria of a thermal imager is its detectability (Detectivity, D*). It is often determined by the ratio

where

• A is the area of ​​the sensitive element of the photodetector,
• ?f is the effective value of the bandwidth (noise).

In the work [4] the author gives a more detailed method for calculating the detectability:

In this formula

• ? – operating wavelength,
• ?, h, c, kB, e – physical constants,
• ?? – range for which the detectability is calculated.
• ? – solid angle.

The NETD parameter can often be found in specifications– Noise Equivalent Temperature Difference – equivalent noise temperature difference. This value is equal to such a difference in the temperatures of the scene and the object, which is assessed by the device as noise. It is introduced to simplify the understanding of the sensitivity of the thermal imager. For example, for a cooled thermal imager with NETD = 20 mK, this means that a body with a temperature of 30.002 ° C will be indistinguishable from the background with a temperature of 30 ° C, while a difference of one degree will be clearly noticeable. Interestingly, a rattlesnake has approximately the same sensitivity as a cooled thermal imager – 18 mK. True, the use of snakes in thermal imaging is not very practical: the distance at which such a «thermal imager» looks is no more than a meter.

The above materials are necessary, but insufficient information for the high-quality design of a surveillance system operating in the infrared range.

_____________________

S. Nikitin, «SKN»