Pyrometric Human Presence Sensor

Pyrometric sensors measure high temperatures. Are humans you wish to detect the presence of on fire? Why not use ordinary PIR sensors? They're much cheaper.

Dear Bill,

Pyrometric doesn't mean flames. It is just a heat measurement technology. PIR is Pyro-Infrared sensor. Anything above 0K emits radiation and there is lot of radiation from human body. We intend to measure hot body in cool environment and cool body in hot environment. In India temperature ranges from -55C to +55C. Human body is at about 37C or 98.6F.

These engineers are trying reverse of motion detection and they are scanning the presence by moving the detector. I think there will be greater noise, but let us see their results in few weeks as they are up their sleeves doing this work. Yes, they are using PIR sensors, Deep-IR sensor, MIDIR sensor and will try anything that can work from 1-10 micron wavelength.

Perhaps water molecules on human skin can also be sensed using absoption bands in 900nm to 2200nm range. These bands are used in remote sensing of moisture on earth.

Well, I've never heard or read pyrometric or pyrometry used in the context of measuring anything but high temperatures where infrared radiation is significant. My unabridged, Webster's Dictionary refers to "high temperature" also. I have also never heard or read PIR where it meant anything other than "Passive Infrared Radiation" rather than Pyro-Infrared, as you suggest. It's a matter of semantics, I suppose.

Digitally store a profile of the thermal image scan and compare to subsequent scans. Any rate of change of blocks greater than a certain value will indicate presence of a body that is warmer or colder. I saw a similar technique used in a digital video system around 20 years ago for surveillance -- anything which changed from last scan was highlighted on screen.

Dear Bill,

Basically what these engineers are using is Pyro-electric sensor. Pyro-metric is the measurement of temperature by radiation emission. How much temperature or how much radiation is subjective.

IR digital camera cost US$30000 range. PIR is much better for students who have US$500 total investment plan. Dual element differential sensor gives differential peak and easily detects human presence even in hot or cold background.

PIR detector is equivalent to single pixel sensor of the IR CCD camera and perhaps can integrate more signal than large number of pixels of small area can do. Pixel size in PIR is about 2mm while in IR CCD it is micron size.

IR focusing lens is very expensive for imaging work but not for crude total light integrating lense used in PIR detectors.

Have you used such detectors?

I have used PIR detectors, but I have not tried to do anything with them out of the ordinary.

You mentioned that they wanted to mechanically scan an area rather than using a static sensor and detecting human motion, virtually yielding the same results, but that your concern is that it may be noisy. With proper isolation of signal from electromechanical device to move the sensor, there should be no discernable noise at all.

Sources of signals for an IR sensor "looking" at a field, whether static or scanned, must also be taken into account, such as electric heaters, refrigeration condensers, etc., to determine whether or not there is motion.

I'm too removed from the problem to be of any help, but it's been interesting.

Thanks for something to think about.

Dear Bill, Scanning the PER detector can become noisy because the background itself is changing and may have different temperature and that difference is not a wanted signal for these engineers. They are actually looking for background - human signal. Noise actually may come from background_1-background_2=noise type signal, which is actually noise and not a signal at all. Stable sensor and stable background can not easily gemerate noise other than normal electronic noise and perhaps wind temperature noise.

When we are looking for very small signal, each thing counts. Temperature of corners will be lot different from mid wall and near door temperatures. Also some area may remain hot where a person was but not present now. Bulbs and other light sources, ventilation and cooling system bound to cause noise due to thermal disturbances.

Engineers have planned an intelligent program to remove information that is noise and not from presence of human. They can not improve the sensor.

It seems to me Capacitive sensors would be cheaper, and just as effictive to sense human presence. They would be stationary, and less prone to failure. A room could be set up with plates set on the walls, and a computer could switch from one set to another, and no two plates have to be a pair. Any two can be paired. with the computer switching the plates randomly,changing each plate to pair with another plate for a pair, it would be possible to exactly locate the person, or persons in any part of the room. In any design the most simple approach is best. For security, set up four computers to monitor the room. Use three all getting the same feed. When any two agree, and the third does not agree, drop the third out of the loop, and swith the forth in. (As long as all three agree, there is no malfunction.) When this happens, notify a technician to replace the bad unit, and rebuild the one removed. This should make a very secure room. Of course; the Walls Cap plates must not be noticible. The plates can be stationary, and set in pairs. set the plates every foot apart, and paired with the Plate in direct opposition. Thus, when one set had senses a person, then on the walls at right angles a cap set there will show a person. Now you have location of the person. Assume the room is set north-south and east-west. When the a north south walls cap pair show a person, the East West walls cap pair will show the same person there om one set of caps. As the person moves, other sets on all the walls will show the person's progression across the room. The east-west pair and the north south pair form an X for the location of the person. I was thinking as I wrote this. I got a better idea. If this is a little fuzzy. One plate is on the north wall, and the other of the pair is directly opposite on the south wall. Same true east west paired cap plates. . Fris.

Dear Fris,

Capacitor sensors made out of metalic plates of adjacent walls can not resolve the number of people present in the room even if you switch the plates, as they are not sharp enough when plate separation is too large. They also become noisy as entire room is a noise source. Plates with large distance can form onlt small capacitance. I agree this sensor can yield good average sensing mechanism. However, it is also going to be sensitive to the material present in the room, moisture. Remember dielectric constant of water is 80 times to the dry air. Capacitance is also going to change that much factor.

Perhaps if capacitor sensor is made on floor then it may work much better unless people are in air or sitting legs up on something.

Thanks for the Idea. I like and greatly appreciate such thinking. Perhaps top roof can be one plate and many small plates on floor can be switches and scanned. I am turning the idea 90 degree now. Looks that may also take care of the people in air and not exactly on floor but somewhat near to the floor. Sounds Good, but I am never so sure of capacitive sensors with plates very far away. Above all that are electrical cable with huge AC signal radiating all over inside this capacitor. I am not sure if any one likes AC power wiring inside a sensor capacitor.

Thanks for the reply. If you use one cap sensor shielded from room, but able to sense room air, you have a bechmark for the calibration of all the cap sensors. I screwed up. I did not know there were to be many in the room. your idea of the floor sensors would be practical, unless someone dropped a coat, or dropped a drink on it, etc. I would use a cross section of Caps wall placed, and IR sensors in the celiings down sensing. Each with a tight focus of about 6" diameter. Set them apart about every 9 inches, so there would be overlap. The room could have a red celophane membrain to obscure the sensors. Set one in the wall as a benchmark, by having it at ceiling height. Sensing across the room. Thus you would have a benchmark for room temperature, and as the heating or cooling comes on or off, the benchmark would correct for it. I think this could be a passive setup to gain all the result you need. Not too complicated. At no time aim the IR sensors at any florescent lit show case. When the lights come off in the case, it cools down and will cause havoc with the IR sensors viewing the case. But if you make sure the case lighting does not shut down you shouod be O. K. Caps and IR sensors could sort out the info you need. Fris.

I agree that floor may have many unwanted things and to play safe some height as starting point is much better idea. Good thinking.

Plan is for Pyroelectric sensor and an Ultrasonic sensor. While Pyroelectric has been planned to see if there is some IR from living organism of the size of commonmen, the ultrasonic sensor is planned to see, how far the person is located from the sensor.

I can easily have an ultrasonic sensor for 12m long room. This sensor will also assure that there is something in the room as it will bhave an idea of a blanck space scanned earlier in the memory. Using ultrasonic sensor one can get very clear picture. Unfortunately I can't use an array due to cost involved.

Shyam; Untrasonic sensing is O.K. but room air motion can trigger false alerts. Best to use IR sensor next to heating or cooling vent to let Ultrasonic know there is air motion. The heating and cooling can also affect the IR sensors. Try Alarm ultrasonic sensors. I worked with them back in the early 60's. As long as the frequency doesn't drift too much, and the air movement is limited, it will function quite good, as long as it is just looking for motion. It will be difficult to sense several at the same time. IR is good equipment, but can give false reads, if there is a case with Balasts for flourscent tubes. When the llights are cut off, the ballasts cool and usually cause the IR to trip. So keep the IR units not focused on Jewelry cases. I worked with IR guns that can be focused to read air temps in ducts. I had a problem of Air Stratification in a drying heating duct. I checked out the air across the duct and found Hot and cold streams within the duct. I welded baffels in the 90 Degree duct, and that mixed the air to get even temp across the Duct, and cured the stock properly. I ramble because I have no idea of your room conditions. I would generally go with the Passive units as much as possible. When set up right, they are greatly reliable. In school, my group set up a Locker as a Capacitive section of a tuned circuit antenna, and it was adjustible to trip the alarm with one person within 20 feet of the locker, to needing five pepple within several inches from it, to trip. Passive sensors, with limited heating or cooling air velocity, and moderate temperature fluctuation, would seem to be the best to me for non intrusive alarms, and most stable. Good luck in your endeavor. Fris. Capacitive sensors work well, unless there are big changes in moisture in the air, but using one as a bench mark Not set to look at people, but just sense the air at ceiling level. will allow the others to function properly and ignore the moisture change Fris..

Hi Fris,

Room air can't affect ultrasound sensing. Range sensing is done by time signal and air velocity for ultrasound does not change by small air flow or wind. Position sensing as accurate to 1mm is possible using ultrasound.

IR range sensing may also be good idea butI need a CCD device and this means high cost. Engineers have plan for US$500 only in all. Even ultrasound ranging professional sensor fro www.turck.com comes at that price.

Sinple iR sensor with quantitative reflectivity monitoring will be with lots of error. Laser ranging sensors is also ruled out for cost and requirement for ultrafast professional logic down to ps accuracy.

Our choices are very limited.

Tomorrow morning engineers will experiment with one Pyro sensor to see the output of the sensor without lense or with pin hole lens. We can discuss after we get this result.

I am sure, your discussion has helped a great deal and all engineers are also reading your communication. They told me this morning about it. We have enough time and sure will build good system in about three months.

Keep writing.

Dr. Shyam,

I had just blindly bought a pyro transistor without data. Since you are in the field of Pyro-measurement, can you provide me with a datasheet for the TP302 (it is written in the body of the transistor).

No info about who making this transistor. It is T0-39 package with flat lense on top and four legs (none of the legs connect to the transistor body physically and electrically).

I did try to search deep into the web, ask the seller too, and no results.

I am in a progress of making a human body heat detector design from the scratch.

I always wlecome to get new perspective of how to detect this human body heat if we have obstruction say walls, ruins, etc.

Warm regards,

Achmad

Perhaps you may have a phototransistor, which may not be pyroelectric device. Phototransistors may work only up to 1000nm while Pyro are meant for 10000nm.

Dr. Shyam,

It Might be Phototransistor, but I will search the data till I get one that able to tell me what it is realy.

Sir, My country Indonesia till now is well known as a smoke exporter to Malaysia, Singapore and Brunei. This is not good, because the smoke is a result of bush fire include forrest fire (wildland fire) done by the unresponsible people.

So, I try to study Pyroelectric devices and would like to use them in the project. By studying this device and found PIR L14G3 (photo transistor) I built Sectorize home security system. I hope I could have the pyro for detecting human body heat. The sensor is place outdoor covering certain distance covering the yard.

I know by using the PIR more mall-trigger can happens. Can you say things for my additional awareness about utilizing both PIR and Pyro for outdoor detecting system?

Dr. Shyam,

I want to design a cellular like system of monitoring wildland fire. I had been assisted by my friend in US, that there is a device from Hamamatsu but it works by 325volts. Can you direct me to the right device?

Thank you Sir.

regards,

Achmad

Human Motion Detectors, which use Pyroelectric sensor are in the market for about US$50. Just buy one.

It is unfortunate that Indonesia is going doen in economy even when you have so many hardworking people around.

I actually met one Atomic Energy Scientist from Indinesia his name Subari Samanto in United Nations IAEA meeting in Santo Domingo, West Indies in 1988. I think at that time your country was much better. You have to help your country for sure.

I think problems with your country are more towards management now. It is rarely that people are responsible. People are often used by others.

I think I am drifting away from what this topic is all about. If you need just one sensor, then I will say buy it. If you want to do business then learn to do it from ready technology first.

Yes Dr. Shyam,

I agree to what you all saying. Like you say, I just trying to do something that will worth for all which is try to finding device and expertise such utilizing pyro technics to early warning bush fire. The pyro hence never been used for such purposed. Anyway, thank you for your reply so far. I remain.

Very Truly yours,

Achmad

For detection of fire from long distance is very difficult. However, you can detect from near by area and by then fire is already there. Smoke comes earlier as it travels much earlier than fire. Hence detect smoke first. Just visibility sensor will do. It will also tell where fire is by looking at smoke signal. People in early times used to communicate using smoke signal. Fire communicates to you by the same means.

In all buildings etc also smoke detectors are first to use. IR comes only when there is full fire.

I am not saying that you need not learn. What I am saying is you don't waste money and use only right type of technology.

Dr. Shyam,

Yes Sir, I agree with what you says. I am now in the middle of searching for a working early fire detection system, devices involved, techniques, etc. Hence, Zero cost for me but my coffee.

Surveillance Tower had been raised and function for long time ago, but not effective due to lack of sophisticated equipment. To detect smoke the only gadget they use is Binocular. It was working but mostly not well due to human error is 99% (random number)take place - thus it is not bad but it is not good. Here, I play the role as requested, either suggesting them to buy products, or built them as per readiness of devices.

For long distance early fire detection I had to conclude that pyroelectric is not the correct selection due to distance. (Correct me if I conclude wrongly on this)

PIR on the other hand, capable of monitoring Long distance with correct lenses. (I had info from my colleague in Canada). They use, PIR, Lenses, Mirror, and frequent move just to find the little differece in measurement. Thus constant slow move gives better result of forrest surveillanve. I depicted they will have problem in motorizing the sensor. (later he says: Correct).

I had wonderful intense contact with the UNOSAT, I will integrate their IR Satellite imaging into my solution. After the earthquake recently on May 27, 2006, They help me a lot for this imaging data. And suddenly, exactly 11 days from the earthquake, I had successfully design an equipment capable of monitoring and give alarm to community of the incoming earthquake. Amazingly, it can detect earthquake event more than 700km from my town. (all the following earthquakes from July-August 2006 in Indonesia was detected and firing the alarm).

I am wondering, might I make same success in the field of wildland fire early warning system. As simple as that.

warm regards,

Achmad

Yes, I fully agree that remote sensing from above is the best way to habdle this task as you have many wavelength to scan. It is all right to sense IR and UV bands in that case as UV comes from intense fire and IR comes from start up fires as well as from intense fires. You can also have an estimate of the fire damage and exact location of the fire. It is sure much better idea.

If you do samething from tower, then you need Germanium lens large size high resolution CCD camera which can map the same parameter with some difficuly as there will be obstacles on the way.

LIDAR is good to sense smoke and smoke contents and hence fire. This can be done at ground level as smoke goes up high and you can point beam directly towards sky.

I think, you should hire experts and your government should be willing to pay for it.

Yes Sir.

I had in mind proposing them the expert assistance. There are almost zero expert here. But, my input to the government somehow dirrect them into the right path already. The expert if I am not wrong came from Canada.

OF course the IR+UV, LIDAR, etc so far proposing neat solution. My role is to open the solution as wide as possible to the government, drive them to staggering or step by step action to be taken, and all these becoming the guidance for making such project, RFP etc.

Anyway, I thank you the sharing Sir. Might I post you later decision making heading just for all engineers in the forum references. Not a promise but I will do my best upon permission from my superior.

Again, I shall conclude the discussion and fully thanking you of your awakening.

Very truly yours,

Achmad

Hello,

I have found your site interesting and hope that you will be able to assist me in a project. I have already installed a DVR (Digital Video recorder) in a police station jail cells. Currently they need to record 100% of the time in the cells to ensure they get recording 100% of the time a person is in the cells. They cannot use a manual switch to turn on recording, the guard might forget to turn it on, or will turn it off to harm the inmate. They cannot use motion detectors, as the sleeping prisioner will not be moving enough to trigger the sensor. They want an automated detection system that will only record when someone is in the cell, but that is 100% automated and fail proof.

My thought is to us a sensor that detects the body heat in the cell, not movement in the cell. When ever there is body heat the recording will happen.

Do you know of a sensor that will do this? Please email me at rbr@patronuslabs.com if you know of such a product.

Thanks,

Rod Russell, CTO, Patronus Laboratories Corporation

Dear Rod Russell,

Pyro-electric sensor and many IR temperature sensors can sense the heat signal. However they can not tell if it is from a man or from a bulb. That needs special sensor which can say for sure that there is a man inside. This requires very accurate temperature sensing near 37C +/-2C which is human body temperature. These sensors are a bit expensive.

If you are monitoring criminals then they can be radio tagged and even if they move you will get radio signal. They can not easily remove a radio tag (saw resonator). this one can tell complete history of a man and hence a better device.

This radio tag you can not do unless law of your contry permits you to do that. It does not harm a persom but sure affects the privacy of a person to much greater extent. He or she may become mentally sick if one knows that they are monitored by a radio tag. This may be a human rights problem.

Fascinating..... so a £3 PIR sensor to detect a heated body emmision isn't suitable?

You want to differentiate between a bulb and a 36 *C human?

The sensors we used in a portable infrared thermmeter were pyro sensors, we used a focused reflector to avoid the costs of germanium and arsenic lens systems.... For infrared a brass coloured reflector is excellent, to exclude other visible frequencies we used a fine plastc film over the opening... with careful setting up we could achieve targets of 2" diameter at 30 feet distance and monitor the temperature to within a few degrees....

Maybe you don't need a scanning pyro sensor just a standard movement PIR together with a pryro temperature measurement sensor to only detect movement within the human bodies range of temperatures???

John.

Dear John,

That is a good idea to use reflector or mirror lens. I am not having data on reflective nature of metals for deep IR in 10 micron range. Let me see if I get some optical data. Can I have a picture of what you have designed? Looks interesting idea. How did you arrange it? Like a salellitle disk antenna amplifier? 10m is a good thing. Most people use thin thermocouple or heat flux sensors and some use Pyro-electric sensor

I will also get one digital Pyroelectric Sensors now to get the 14-bit IR intensity digitized data rather than analog version. I am still worried about the lens as I don't have proper source for it. I need telescopic IR lens such that I can collect IR from small zone over a distance.

Dear Shyam,

For 10um or longer wavelengths brass is a superb reflector, although for cost reasons we chose a metallic plated plastic mirror for the focussed reflector.

The construction is identical to a reflecting Newtonian telescope, the pyrometer sensor was positioned at the focus of the mirror, a thin plastic film was placed over the entrance.... the measuring tempersture range was from below zero to many hundreds of degrees Celcius, emissivity could be set for shiney none black body targets etc....

This was designed in 1976 at 'Kane May' and featured on 'Tomorrow's World' TV programme - A programme highlighting new designs / inventions made etc...

The demonstration we set up in the TV studio was to show how even with bright studio lighting and the background temperature the thermometer could spot the minute temperature difference on a wall made by having a candle burning on the other side, and so identify the position of the candle...

It was supposed to show how it could be used in a search and resue application looking for human bodies trapped in rubble etc...

Kane May Ltd, is sadly no longer here, but the USA version was made by 'Land Co.' and is virtually the same design.

John.

Brass is pretty cheap in India but making a mirror out of it may be a problem. I think the torch light reflector may also work. I will say, good idea and you did it 30 years ago. That was my starting point. I did work with RDX explosives at that time for creating explosion under water in nuclear reactor pipes to test their strength to withstand accidental pressure pulse. We take lots of care in nuclear technology design.

I wan't in hobby ideas for many years as one or the other serious nuclear electronics was to be developed. I designed small MCA for 8192 channels that could be placed in hand while ORTEC was making huge box type. I think what I designed is still working after so many years. Our engineers used it all the time and never put it off even for the last 28 years. I think now they might have thought or removing it. Not that it is not working but they may like to work with new computers. I have designed to work with character terminals. There wasn't much of graphics at that time in 1976.

Let me see if I can try this new suggested idea.

From my understanding, your research focuses on if there is human and number of them, are you interested in human's position in the scanning area or not?

Idea is to sense the number of human presence in the room on dynamic measurement bases. Problem with this type of measurement is the distribution of temperature in the room for different objects. Motion sensors can easily differentiate change in the status for stationaly background. Here we are altering the background itself during scan. Perhaps by storing the background signal under multiple scan can help a bit better.

Other use highly expensive laser scanner for the distance and is sure more accurate method. Our idea is to make is very cost effective. This part of the research is to examine the way our thermal sensors will respond to the situation. Other options are not closed but we need more ideas to experiment with. We are aware of the problem and are looking for a solution within the accepted sensor type and relatively noisy information.

Please send the technical details of the system as sensor type range etc along with the cost. Mail the Details to atul_shitkande@relbio.com .

Regards,

Atul Shitkande

Reliance Life Sciences Pvt. Ltd

Mumbai, India

Can my numbers from my web site.

HELLO I AM MAKING THE PROJECT FOR DETECTION OF HUMAN.....................

IS THERE ANY (((((SENSOR)))))) PIR OR ANY OTHER ANY ONE RECOMMEND ME,.......,,,,,,,,,,,

BUT IT MUST DETECT NOT ONLY THE MOTION OF THE HUMAN BODY BUT ALSO DETECT THE STATIC HUMAN BODY ,,,,,,,,,,,,,,,,,,,, WHICH IS NOT MOVING,,,,,,,,,,,,,,,,,,,,,,,,,

ANY ONE HELP ME OUT,.................................................

........................................................................

Look for PIR sensor in local market. Search for information on its application on the net and do some home work if that is your project.

hi..................................................

i got stuck in making the decision for the sensor in the designing process in our city .I FOUND SOME SENSOR ON NET WHICH ARE HELPFULL BUT NOT AVALIABLE IN MARKET THAT ARE (IRA E700, IRA E900, RE200B).

The following are the sensor avalible IN OUR CITY. can it be possible to detect the (PRESENCE OF HUMAN) with the help of the following sensor..................

**************************************************************

FIRST SENSOR

OMRON: E3JK-DS30M1

10U8M

**************************************************************

SECOND SENSOR

SCHEIDOR ELECTRIC TELECANIQUE: XUXIARCNT16

**************************************************************

THIRD SENSOR

SUNX NX-52M-S1

*************************************************************

IF THEY ARE USE FULL GIVE ME SUGGESTION OR ANY OTHER WAY TO DETECT THE PRESENCE OF HUMAN.

IN MY PROJECT WE HAVE TO DETECT AN HUMAN AND SWITCH ON THE LIGHT UPTILL THE HUMAN IS PRESENT AND WHEN THE HUMAN IS GONE THE LIGHT WILL AUTOMATICALLY SWITCH OFF.THE PROBLEM TO ME IS THAT IF I GOT A SENSOR THAT SENSE THE PRESENCE OF HUMAN AND GIVE ME SIGNAL THAT WITH THE HELP OF CONTROLLER I WILL SWITSH ON OR OFF THE LIGHTS..............

Hi, i just wanted to know if you were able to find a good sensor for your application.

Dr. Shyam,

It has been a long time since you last posted any comments in this thread. I am interested in your research findings. It is a very interesting topic and I like to know if research finds it feasible.

Thanks,

This was a project by a team of engineers in education and they completed it long ago. It was not a commercial project for me.

Dr. Shyam,

I appreciate your reply. Honestly, I was not expecting that you are still monitoring this site.

Is there a way to get the results or get in touch with the students?

Thanks again for your time.

Dear Shaym I am interested at the findings done by your team of engineers. Is there a way for me to contact them? Thanks.

i am deploying home automation i want to know human presences in bed room while he sleep also this can not achieve through pir plz suggest me what to do