IR thermometer vs thermal imager

[chirhone]

Isn't it both use the same exact principle? An infrared thermometer versus a thermal imager? Why is it that the thermal imager is not for body temperature (as indicated below) while the infrared thermometer is for body temperature. I think they both scan the same blackbody of the physical body.

 

[Drakkith]

Not having personally used a thermal imager before, this is my best guess:

The thermometer is designed for reading temperatures and is calibrated as such. Thermal imagers are not. While they do detect thermal radiation, they are most likely not calibrated for detecting small changes in temperature, and if they are, they aren't designed to measure the actual temperature. They instead show you the temperature relative to an object's surroundings in the form of colors and shades. At best, you might be able to tell if someone is warming than someone else if you have both of them in the camera frame at the same time.

Along the same vein, a cell phone camera cannot be used to measure the exact intensity of a light source without knowing a great deal about the detector's sensitivity, filter sensitivity, gain settings, and properties of the light source. These are not normally available to the average user.

And, in the event that thermal imagers are able to reliably detect a fever, the company still doesn't want its products being used that way for liability purposes. I'd be willing to bet that they don't have the product certifications required for medical use.

 

[anorlunda]

I think the operative part of that warning is this

Bosch will not be held liable ...

But think of the small temperature difference between a fever at 101F (38.3C) and 98.6F (37C). The device may not be that accurate. It is entirely possible that the accuracy and repeatability are in the range of 2-3 degrees C, or 4-7 degrees F or worse.

 

[Dr_Nate]

I think they both scan the same blackbody of the physical body.

They aren't blackbodies. Various objects have different emissivities.

The infrared thermometer is going to be calibrated to the emissivity of human skin in some wavelength range that is not sensitive to melanin content. The measuring spot will also have to not have large variations in temperature over different body types (think adipose tissue).

Industrial thermal imagers are most definitely not calibrated that way.

 

[zoki85]

Various objects have different emissivities.

The infrared thermometer is going to be calibrated to the emissivity of human skin in some wavelength range that is not sensitive to melanin content. The measuring spot will also have to not have large variations in temperature over different body types (think adipose tissue).

Industrial thermometer I use at work is certainly not intended for measurement of human skin temperature. However it can measure both, temperature of hot metal pipe (~100 °C) and temperature when pointed at my forehead 36.5-37 °C. I guess precision is at least ± 0.5 °C

 

[Dr_Nate]

Industrial thermometer I use at work is certainly not intended for measurement of human skin temperature. However it can measure both, temperature of hot metal pipe (~100 °C) and temperature when pointed at my forehead 36.5-37 °C. I guess precision is at least ± 0.5 °C

You might have gotten lucky and have similar emissivities for those two things in the wavelength range measured. Check out the temperature measurements on this Leslie cube here at around 4:20: .

Also, those temperature are good for core temperatures. Skin temperatures are lower than core temperatures. You'll need another calibrated measurement to know how accurate your measurements are.

 

[zoki85]

One African girl works in a restaurant of my company building. I'll ask her if I can take measurement of her forehead

 

[sophiecentaur]

One African girl works in a restaurant of my company building. I'll ask her if I can take measurement of her forehead

I wonder whether you would get a different reading. The Chrominance of black and white skin is more or less the same* so I conclude the black pigment is pretty much neutral density. There may be more or less energy emitted but the spectrum will probably be similar.

* I don't have a citation for this but I worked in Colour TV technology for a while and it was 'common knowledge' amongst my colleagues, who were frequently looking at the phase of the PAL Chrominance signal. You can always do an experiment yourself if you have a photo containing black and white skinned faces and you look at the ratios of RGB in both cases. (That Colour Dropper tool is very handy for all sorts of things.) Something to do when you are self - isolating.don

Edit: I don't actually know this but it's likely that the pigment in skin is actually there to deal with UV radiation so the emissivity at the IR wavelengths could well be much more similar for all colours of skin. Whether the colour makes a lot of difference is open to question, bearing in mind that the skin of polar bears is actually black and you'd have thought that they would require as little heat loss as possible from the skin. A puzzle.

 

[chirhone]

I bought the popular foreheard ir thermometer to compare with my thermal imager.

So far. They are so close as to be almost similar. And I have the following questions.

1. It seems a thermal imager can actually detect the actual temperature and not just show you the temperature relative to an object's surroundings in the form of colors and shades, right? What principle in blackbody and emissivity literature wherein an object actually emits the true vibrations corresponding to the actual temperature of the object?

2. Many forehead thermometer has Surface and Body mode. The Surface mode is lower in temperature. I don't think it's just lowering the emissivity. Because in a thermal imager, lowering the emissivity setting will make the temperature higher, not lower.

This is the settings in the IR thermometer.

This is pointed to my forehead with Body mode.

This is pointed to my forehead in Surface mode.

What do they have in mind when they lower the temperature when using Surface mode? Its not lowering the emissivity setting which can increase the temperature in a thermal imager.

 

[russ_watters]

What principle in blackbody and emissivity literature wherein an object actually emits the true vibrations corresponding to the actual temperature of the object?

That's a very poorly constructed sentence and is hard to understand. If you're just asking how an infrared thermometer works, you can google for that. Here's a link I quickly found:
https://www.sensortips.com/temperature/infrared-temperature-sensor/

2. Many forehead thermometer has Surface and Body mode. The Surface mode is lower in temperature. I don't think it's just lowering the emissivity. Because in a thermal imager, lowering the emissivity setting will make the temperature higher, not lower.

It's probably an estimate of how much warmer your body is than the surface of your forehead.

 

[Drakkith]

1. It seems a thermal imager can actually detect the actual temperature and not just show you the temperature relative to an object's surroundings in the form of colors and shades, right? What principle in blackbody and emissivity literature wherein an object actually emits the true vibrations corresponding to the actual temperature of the object?

Let's clear a few things up. First, objects don't emit vibrations. They emit electromagnetic waves, which are not physical vibrations but alternating electric and magnetic field vectors. Second, IR imagers and thermometers do not detect temperature. They detect IR radiation and use that to compute the temperature. A typical IR thermometer works as follows:

Infrared thermometers usually use a lens to focus infrared light from one object onto a detector called a thermopile. The thermopile absorbs the infrared radiation and turns it into heat. The more infrared energy, the hotter the thermopile gets. This heat is turned into electricity. The electricity is sent to a detector, which uses it to determine the temperature of whatever the thermometer is pointed at. The more electricity, the hotter the object is.

Source: https://sciencing.com/infrared-thermometers-work-4965130.html

As for IR imagers detecting temperatures, I admit that my previous guess is likely wrong. See here:

Inside of a thermal camera, there are a bunch of tiny measuring devices that capture infrared radiation, called microbolometers, and each pixel has one. From there, the microbolometer records the temperature and then assigns that pixel to an appropriate color.

Source: https://www.howtogeek.com/294076/how-does-thermal-imaging-work/

So it appears that each pixel in a thermal imager is essentialy its own IR thermometer.

2. Many forehead thermometer has Surface and Body mode. The Surface mode is lower in temperature. I don't think it's just lowering the emissivity. Because in a thermal imager, lowering the emissivity setting will make the temperature higher, not lower.

I'm betting that all that's happening is that the thermometer adds a small amount to the temperature when in body mode to make up for the fact that skin temperature is lower than core body temperature.

 

[chirhone]

That's a very poorly constructed sentence and is hard to understand. If you're just asking how an infrared thermometer works, you can google for that. Here's a link I quickly found:
https://www.sensortips.com/temperature/infrared-temperature-sensor/

It's probably an estimate of how much warmer your body is than the surface of your forehead.

The Surface mode is used for measuring surface of an object and not the body.

The Body mode is for measuring the human body.

I'm trying to understand how the algorithm of Surface vs Body mode works. When it's set to Surface mode and you aim it at surface of say the wall, the temperature is lower than the body. I'm comparing to my thermal imager to understand the algorithm and what emissivity adjustments they were doing.

 

[chirhone]

Let's clear a few things up. First, objects don't emit vibrations. They emit electromagnetic waves, which are not physical vibrations but alternating electric and magnetic field vectors. Second, IR imagers and thermometers do not detect temperature. They detect IR radiation and use that to compute the temperature. A typical IR thermometer works as follows:
Source: https://sciencing.com/infrared-thermometers-work-4965130.html

As for IR imagers detecting temperatures, I admit that my previous guess is likely wrong. See here:
Source: https://www.howtogeek.com/294076/how-does-thermal-imaging-work/

So it appears that each pixel in a thermal imager is essentialy its own IR thermometer.
I'm betting that all that's happening is that the thermometer adds a small amount to the temperature when in body mode to make up for the fact that skin temperature is lower than core body temperature.

The Body Mode is for measuring 35 degree C to 43 C only. Surface Mode is for measuring objects like water or wall. See the following. I can't understand chinese but can read the numbers. Anyone understand them?

In our worldwide Lockdown. We must understand these things thoroughly because it will be with us for weeks or months.

 

[Dr_Nate]

1. It seems a thermal imager can actually detect the actual temperature and not just show you the temperature relative to an object's surroundings in the form of colors and shades, right? What principle in blackbody and emissivity literature wherein an object actually emits the true vibrations corresponding to the actual temperature of the object?

The image of a thermal imager is just going to give you a measure of the radiant power that the object is emitting in the wavelength range of the imager. (It is possible that it a measure of the photon count, but I doubt that). When it measures the temperature, you must make assumptions about the emissivity.

The actual spectrum of radiation emitted by an object is a convolution of blackbody radiation and emissivity (as a function of frequency or wavelength). This is how I think about the origin of emissivity in a mathematical sense: electric polarization $\rightarrow$ any pair of optical constants ($n$ & $k$, $\epsilon_1$ & $\epsilon_2$, $\sigma_1$ & $\sigma_2$, etc.) $\rightarrow$ reflectivity (R) $\rightarrow$ emissivity = 1 - R.

 

[chirhone]

Anyone can understand chinese? What is this definition of Body vs Surface? I saw in youtube Surface is used for objects and Body for human body. My thermal imager doesn't have these settings. It can see both. Why a thermopile needs separate mode than a bolometer (microbolometer used in thermal imager)? :

 

[russ_watters]

The Surface mode is used for measuring surface of an object and not the body.

The Body mode is for measuring the human body.

I'm trying to understand how the algorithm of Surface vs Body mode works. When it's set to Surface mode and you aim it at surface of say the wall, the temperature is lower than the body. I'm comparing to my thermal imager to understand the algorithm and what emissivity adjustments they were doing.

They aren't adjusting emisssivity. You are making this too complicated. Again: Body mode is an estimate of how much warmer the inside of your body is than the surface of your forehead.

 

[Dr_Nate]

Why a thermopile needs separate mode than a bolometer (microbolometer used in thermal imager)?

The bolometers I have used need to be cooled to a very particular temperature to be in calibration. You aren't going to be able to get temperature control in a small handheld device like an IR thermometer. It would require a Peltier cooler and a good power supply. This wouldn't be cheap or light. The solution then is a thermopile.

Why a bolometer in an thermal imager? It's likely because they respond fast to incident radiation, hence the ability to make a video feed.

 

[chirhone]

The bolometers I have used need to be cooled to a very particular temperature to be in calibration. You aren't going to be able to get temperature control in a small handheld device like an IR thermometer. It would require a Peltier cooler and a good power supply. This wouldn't be cheap or light. The solution then is a thermopile.

Why a bolometer in an thermal imager? It's likely because they respond fast to incident radiation, hence the ability to make a video feed.

My microbolometer is handheld on the right. The left is the ir thermometer.

This is when i aimed the right on an electric fan.

 

[chirhone]

They aren't adjusting emisssivity. You are making this too complicated. Again: Body mode is an estimate of how much warmer the inside of your body is than the surface of your forehead.

If the above is the case. No problem. But using the thermal imager on a friend. The surface really measures 36.6 C and not 35.5 C.

Hope others who own an ir forehead thermometer presently used in checkpoints all over the world can verify the meaning of Body mode. Or if someone has english manual of it.

Note: a thermal imager like the above doesn't have Body mode. It measures all surfaces only. Only an ir forehead thermometer has Body and Surface mode.

 

[Drakkith]

@chirhone I don't understand your confusion. Skin temperature is less than core body temperature, therefor the thermometer must either be designed to solely measure body temperatures, in which case it will always adjust the raw measurement upwards in some fashion, or it must include settings that allow for the measurement of both human bodies and environmental surfaces. Obviously your thermometer is the latter type.

 

[chirhone]

@chirhone I don't understand your confusion. Skin temperature is less than core body temperature, therefor the thermometer must either be designed to solely measure body temperatures, in which case it will always adjust the raw measurement upwards in some fashion, or it must include settings that allow for the measurement of both human bodies and environmental surfaces. Obviously your thermometer is the latter type.

The FLIR thermal imager measures in raw mode. It is for electrical and constructions. But when aimed at a human forehead. It measures 36.6 C which is identical to the smaller IR thermometer unit in Body mode. This gives me idea the Body mode is measuring human body at the surface. And Surface mode is for inanimate objects. In fact most demonstration in youtube explains it that way.

About surface and core temperature of body. I know core temperature is higher but i don't think the ir thermometer is compensating for it. Those who actually own one pls verify. Thanks.

 

[Drakkith]

The FLIR thermal imager measures in raw mode. It is for electrical and constructions. But when aimed at a human forehead. It measures 36.6 C which is identical to the ir thermometer in Body mode. This gives me idea the Body mode is measuring human body at the surface. And Surface mode is for inanimate objects. In fact most demonstration in youtube explains it that way.

Well, I admit that I may be completely wrong in my understanding, and without more knowledge in this area I'm going to step out of the conversation for now. Good luck in finding an answer to your questions.

 

[sophiecentaur]

Isn't it necessary to be looking at the Spectrum of the emitter radiation? That would suggest the amplitude of radiation at at least two different wavelengths is necessary. Just measuring the energy flux would depend on distance and perhaps area of the emitter.

 

[Drakkith]

Isn't it necessary to be looking at the Spectrum of the emitter radiation? That would suggest the amplitude of radiation at at least two different wavelengths is necessary. Just measuring the energy flux would depend on distance and perhaps area of the emitter.

Yes and no. The sensor/detector receives radiation from some solid angle in front of it, forming an incoming cone that is projected onto the sensor. As long as the 'base' of the cone is completely filled by the surface that you are wanting to measure then distance is irrelevant. This is just like how the Sun (or any other star) is the same brightness per unit of area no matter your distance from it. Moving further away doesn't change the brightness of the unit areas, it just makes the Sun smaller, resulting in fewer unit areas overall.

However, if the base of your cone is larger than your surface, then you run into problems. For example, if you try to measure the temperature of a lit match head from 50 feet away, almost all of the cone falls upon the background, and you will get a lower temperature than you should (assuming the background is cooler than the match head of course).

In other words, if your surface fills the field of view projected onto the detector, then it's fine. If the surface is smaller than the field of view you will get a lower value than you should.

That's my understanding at least.

 

[chirhone]

There is high probability russ_walters and Drakkith were right about Body mode being software adjusted and meant for internal body temperature.

The following is the IR Thermometer at left compared to the Thermal Imager at right. The IR thermometer is set at Surface mode. Notice the reading is lower than the thermal imager when they have the same target.

The thermal imager shows higher reading. Reflecting on it (pun unintended). Notice the emissivity setting of the thermal imager is 0.95. The reading will make sense if the IR thermometer is set to 0.99 emissivity, making the displayed temperature lower.

I want to test if the IR thermometer indeed has setting of 0.99 emissivity, any ideas how to test it?

Also I read 100 Celsius temperature corresponds to the peak near 10 micron wavelength. So does this mean all thermal imagers in the world show the same temperature given the target is the same? So the relationship of wavelength from say 7.5 to 14 micron indeed has a tight relationship to temperature and not arbitrary? Example. The sun peaks at yellow light at temperature near 6000K. This is strictly true for much lower temperature, the peak?

How does the curve looks like at say 36 Celsius? Any ideas?

 

[Dr_Nate]

I want to test if the IR thermometer indeed has setting of 0.99 emissivity, any ideas how to test it?
How does the curve looks like at say 36 Celsius? Any ideas?

I've read that the emissivity of skin in the IR is above 0.98, so that is consistent what you have found.

You want to look up Wien's displacement law. You can also try out this simulation of a blackbody spectrum to get a feel for how it changes with temperature.

 

[sophiecentaur]

There is high probability russ_walters and Drakkith were right about Body mode being software adjusted and meant for internal body temperature.

Medics tend to use the 'in-ear' type, which would give a temperature nearer to body core temperature.

As long as the 'base' of the cone is completely filled by the surface that you are wanting to measure then distance is irrelevant.

That makes sense and the consequence is that an Imager (Looking at a whole scene) would only indicate relative temperatures of objects at a given distance. Fair enough for nearly all uses.

 

[chirhone]

I've read that the emissivity of skin in the IR is above 0.98, so that is consistent what you have found.

You want to look up Wien's displacement law. You can also try out this simulation of a blackbody spectrum to get a feel for how it changes with temperature.

This is very useful. Thanks.

This is the blackbody spectrum at 300k or about 27 Celsius.

This is the peak at 350 K or about 76.85 Celsius

The peaks are so closed together. How does a thermal imager distinguish between a wavelength of say 10 microns vs 8 microns or between 1 Celsius or so? A microbolometer has a range of 7.5 microns to 14-microns wavelength. There are no obvious peaks corresponding to those wavelengths.

Unless it is detecting the intensity of the overall radiation and not the individual peaks? More intensity more radiation and the microbolometer has to be calibrated?

 

[chirhone]

Yes and no. The sensor/detector receives radiation from some solid angle in front of it, forming an incoming cone that is projected onto the sensor. As long as the 'base' of the cone is completely filled by the surface that you are wanting to measure then distance is irrelevant. This is just like how the Sun (or any other star) is the same brightness per unit of area no matter your distance from it. Moving further away doesn't change the brightness of the unit areas, it just makes the Sun smaller, resulting in fewer unit areas overall.

However, if the base of your cone is larger than your surface, then you run into problems. For example, if you try to measure the temperature of a lit match head from 50 feet away, almost all of the cone falls upon the background, and you will get a lower temperature than you should (assuming the background is cooler than the match head of course).

In other words, if your surface fills the field of view projected onto the detector, then it's fine. If the surface is smaller than the field of view you will get a lower value than you should.

That's my understanding at least.

Yesterday I tried to understand these by googling a lot. An IR thermometer has a large cone, imagine it has one pixel only covering the scene or skin. Whereas a Thermal imager has many tiny cones corresponding to each pixel.

Now imagine you are aiming each of them to a surface (such as the wall). The IR thermometer with one large pixel will scan the temperature of the wall versus the Thermal imager with many small pixels scanning the temperature of the wall. Could this result in some changes in the temperature between them since the intensity of each differs? Or the same?

(Btw.. you are probably right about the Body Mode using software to increase the temperature a bit to account for internal body temperature. In the above. I was only considering surface temperature for both the IR thermometer and thermal imager. We are not talking about Body Mode anymore).

 

[sophiecentaur]

Now imagine you are aiming each of them to a surface (such as the wall). The IR thermometer with one large pixel will scan the temperature of the wall versus the Thermal imager with many small pixels scanning the temperature of the wall. Could this result in some changes in the temperature between them since the intensity of each differs? Or the same?

In the end, if you do not want image resolution, the sum of the outputs from a number of small sensors is effectively the same as the total output from a single sensor of the same area so there's no point. If you need to detect a small power flux then it may be cheaper to have a single big sensor and a body thermometer can expect a known amount of power flux over the normal range of temperatures within the proper distance. For a thermal imaging camera you can get the advantage of using a lens which will have more light gathering power and this increase sensitivity (i.e. signal to noise ratio).

 

[chirhone]

In the end, if you do not want image resolution, the sum of the outputs from a number of small sensors is effectively the same as the total output from a single sensor of the same area so there's no point. If you need to detect a small power flux then it may be cheaper to have a single big sensor and a body thermometer can expect a known amount of power flux over the normal range of temperatures within the proper distance. For a thermal imaging camera you can get the advantage of using a lens which will have more light gathering power and this increase sensitivity (i.e. signal to noise ratio).

How about signal to noise ratio? In SLR cameras, the sensors are big and images are much cleaner than that of say smartphone cameras. So doesn't the same comparison hold for the ir thermometer and imager with the latter getting a bit few fluxes and can this lower the temperature?

 

[chirhone]

How about signal to noise ratio? In SLR cameras, the sensors are big and images are much cleaner than that of say smartphone cameras. So doesn't the same comparison hold for the ir thermometer and imager with the latter getting a bit few fluxes and can this lower the temperature?

Or to reverse it. The FLIR imager is $2500, the IR thermometer is only $20. So can the much expensive electronics and better signal to noise ratio of the FLIR create higher temperature than the china made generic thermopile sensor used in the ir thermometer when both aim at the wall of uniform texture or detail and same emissivity?

 

[sophiecentaur]

How about signal to noise ratio? In SLR cameras, the sensors are big and images are much cleaner than that of say smartphone cameras. So doesn't the same comparison hold for the ir thermometer and imager with the latter getting a bit few fluxes and can this lower the temperature?

Size of array, size of pixel and other factors (quantum efficiency etc.) are all relevant. In the end, thought, it's total amount of radiation that gets to the sensor (/sensors) that counts. Big sensor arrays are another matter but the image size in an entry level IR imager will be pretty small cos you pay for area in image arrays. We're not being precise enough for a definite conclusion here but, for an uncooled sensor, there won't be a vast difference in SNR, for a given (total) area and cost.

Can you explain what you mean about "lower the temperature".

 

[chirhone]

Size of array, size of pixel and other factors (quantum efficiency etc.) are all relevant. In the end, thought, it's total amount of radiation that gets to the sensor (/sensors) that counts. Big sensor arrays are another matter but the image size in an entry level IR imager will be pretty small cos you pay for area in image arrays. We're not being precise enough for a definite conclusion here but, for an uncooled sensor, there won't be a vast difference in SNR, for a given (total) area and cost.

Can you explain what you mean about "lower the temperature".

Oh. I realized fluxes can be numerous or few but it is independent of the wavelength, so i guess it won't accept the temperature if more fluxes reach your sensors. Also:

https://en.wikipedia.org/wiki/Wien's_displacement_law

"Wien's displacement law states that the black-body radiation curve for different temperatures will peak at different wavelengths that are inversely proportional to the temperature. The shift of that peak is a direct consequence of the Planck radiation law, which describes the spectral brightness of black-body radiation as a function of wavelength at any given temperature."

 

[sophiecentaur]

Oh. I realized fluxes can be numerous or few but it is independent of the wavelength, so i guess it won't accept the temperature if more fluxes reach your sensors. Also:

https://en.wikipedia.org/wiki/Wien's_displacement_law

"Wien's displacement law states that the black-body radiation curve for different temperatures will peak at different wavelengths that are inversely proportional to the temperature. The shift of that peak is a direct consequence of the Planck radiation law, which describes the spectral brightness of black-body radiation as a function of wavelength at any given temperature."

You are confusing me now. What is the bit about "more fluxes"? There seems to be a problem in translation??

Also, Wien's displacement law can only be used if there is spectral analysis of the incident radiation. A post higher up suggests that the resolution could be difficult to resolve the spectrum of a body at 300K with a simple device.