Early heat seeking missiles use a single sensor. The image is reflected similar to a cassegraine telescope with the reflecting mirror replaced with a spinning mirror that has a triangular modulation window on it. The eyepiece is replaced with the sensor. Your image would be round and the triangular window pattern changes the pulse width modulation of the signal depending on how far the target is from the center. and of course sensing mirror position wil give 360 degree position around yr img.
I don't think you'd get much extra speed by mirror scanning as scan speed will be limited by the thermal mass of the sensor. I wonder if using a chopper might help - removes the DC offset issue of the high-gain amplifier. All the thermal imagers I've see use choppers. BTW Melexis also make thermopile sensors.
Hah, I always wanted to do this myself! Look into compressive imaging (google "single pixel camera"). The proof of concept also used a DMD matrix using a similar setup as your diagram, but in this case the matrix was randomised for each exposure (ie per sensor measurement). Once you take enough samples, you can use the series of samples to mathematically reconstruct the final image via a computer.
You'll want to check out the Melexis MLX90620 for your camera project. It's a 16 x 4 thermopile detector for about $80. pre-digitized and easily read with an Arduino or other microcontroller. Step and stare with that to build up your images.
It could be a digital IR sensor too. Simular to the MLX90614 family IR sensors from Melexis Technologies. They are expensive but they do have an I2C interface.
Are mirrors reflective to IR? You could gut the spinning mirror out of an old laser printer for scanning 'X' in front of the lens... then maybe a simple voice coil (e.g. even a gutted speaker driven by an op-amp) for scanning 'Y' by, I don't know, physically tilting the lens or sensor. Heh. I just had a thought... you should ask Jeri. I bet she'd have some brilliant ideas (does she have any other kind?).
Have you tried to print a raster pattern with crystal paint for better resolution? Great idea. Hope you will accomplish you thermal imaging project with success.
If you coat the film just cut the film into pieces or pixels that can be thermally isolated and collect the data from each one individually or in succession (serialization).
Isn't there a problem with reflectivity of many materials in far IR? I'm just guessing, because many materials, "shiny" in visible light (gold, silver), are bad reflectors in UV. Maybe there's something similar in IR.
Because FLIR came out with the Lepton. Now anyone can build their own thermal camera for like $100. Not to mention the other cheap thermal cameras on the market like the Seek Thermal and the Infiray P2 Pro now.
I'm trying to figure out how to re calibrate these. I jumped the cal pins and when you press the c/f button "cal" show up on the screen but I don't know what to do from there.
You would have to know the exact calibration routine the company uses on the production line. Without knowledge of what temperatures and in what order to present them, you'll just destroy the factory calibration by starting the cal routine.
May please u help me to give hint for calibration, I have one (model: yk-irt2 ) it is showing around half of the real temperature, there is nothing mention in the manual for calibration, I open it, the sensor is three wire and seems connected directly to M.processor, on PCB there is and EPROM, and a place for 8pin IC may be opamp but not used. I checked all the resisters and transistors, found Ok. I tried different tricks to go to calibration or setup mode but couldnt. Now days due to COVID-19 the scanners are short in the market and price goes very high > 250USD hoping to get some help from you, as company is not responding
It would have a calibration routine where, on the production line, each device is pointed at a black body that is adjusted to be as close as they can reasonably get to an exact temperature (probably something like 0 degrees C, 20 degrees C, etc up to 100 degrees C) which the on-board processor can then do a basic linear regression to help offset differences in individual sensors, assuming they are roughly linear. It's possible they do many more temperatures than that, or fewer that are further between. And yeah, the exact voltage at each temperature would be stored in that EEPROM. Without knowing exactly how they do their calibration routine, it'd be pretty difficult to re-calibrate yourself. Theoretically you could probably get a more accurate device if you did have the time and materials to make very precise temperature targets.
Sort of, but there's a limit with fresnel lenses. If you want to really "zoom in" with a thermal camera you'd need some pretty expensive silicon optics.
i think you can build one using a webcam changed to inflared (use negative photo filter) and a software that mesure distance and the contrast and density of the pixels (make a simple mathematic formule to keep this information always available) calibrate all things and you have a webcam that read distance and temperature with incredible accurancy✾☄
Early heat seeking missiles use a single sensor. The image is reflected similar to a cassegraine telescope with the reflecting mirror replaced with a spinning mirror that has a triangular modulation window on it. The eyepiece is replaced with the sensor. Your image would be round and the triangular window pattern changes the pulse width modulation of the signal depending on how far the target is from the center. and of course sensing mirror position wil give 360 degree position around yr img.
I don't think you'd get much extra speed by mirror scanning as scan speed will be limited by the thermal mass of the sensor. I wonder if using a chopper might help - removes the DC offset issue of the high-gain amplifier. All the thermal imagers I've see use choppers.
BTW Melexis also make thermopile sensors.
Hah, I always wanted to do this myself! Look into compressive imaging (google "single pixel camera"). The proof of concept also used a DMD matrix using a similar setup as your diagram, but in this case the matrix was randomised for each exposure (ie per sensor measurement). Once you take enough samples, you can use the series of samples to mathematically reconstruct the final image via a computer.
You'll want to check out the Melexis MLX90620 for your camera project. It's a 16 x 4 thermopile detector for about $80. pre-digitized and easily read with an Arduino or other microcontroller. Step and stare with that to build up your images.
Mlx90640 is about to come out
I think that the set of six pads is a programming header. Vpp usually stands for "programming voltage".
I think the reason why you see those pins change when you press buttons is that they multiplex the programming pins with button inputs.
very informative, thank you.
Muchas gracias por tu video, me dio las pautas para poder armar mi propia pistola de temperatura en base a ese sensor
This project will now have global demand, for a pandemic we are into.
The discrete transistors can be set up in an amplifying configuration, avoiding the need of an omp-amp.
It could be a digital IR sensor too. Simular to the MLX90614 family IR sensors from Melexis Technologies. They are expensive but they do have an I2C interface.
Are mirrors reflective to IR? You could gut the spinning mirror out of an old laser printer for scanning 'X' in front of the lens... then maybe a simple voice coil (e.g. even a gutted speaker driven by an op-amp) for scanning 'Y' by, I don't know, physically tilting the lens or sensor.
Heh. I just had a thought... you should ask Jeri. I bet she'd have some brilliant ideas (does she have any other kind?).
What the focusing black tube called, is that what helps focus the beam for distance?
Have you tried to print a raster pattern with crystal paint for better resolution? Great idea. Hope you will accomplish you thermal imaging project with success.
Could that chip be a PSoC from cypress?
"This is patented of course like every other idea in the world"
+10^6
There's a little pocket laser projector that uses a fairly large "DMD" to send laser beams around.
If you coat the film just cut the film into pieces or pixels that can be thermally isolated and collect the data from each one individually or in succession (serialization).
Isn't there a problem with reflectivity of many materials in far IR? I'm just guessing, because many materials, "shiny" in visible light (gold, silver), are bad reflectors in UV. Maybe there's something similar in IR.
Looks like you gave up on thermal imagery. May i ask why?
Because FLIR came out with the Lepton. Now anyone can build their own thermal camera for like $100. Not to mention the other cheap thermal cameras on the market like the Seek Thermal and the Infiray P2 Pro now.
************* جزاك الله خيرا والله اكبر ***********
Great Job Well done Keep it up Thanks.
I'm trying to figure out how to re calibrate these. I jumped the cal pins and when you press the c/f button "cal" show up on the screen but I don't know what to do from there.
You would have to know the exact calibration routine the company uses on the production line. Without knowledge of what temperatures and in what order to present them, you'll just destroy the factory calibration by starting the cal routine.
May please u help me to give hint for calibration, I have one (model: yk-irt2 ) it is showing around half of the real temperature, there is nothing mention in the manual for calibration, I open it, the sensor is three wire and seems connected directly to M.processor, on PCB there is and EPROM, and a place for 8pin IC may be opamp but not used. I checked all the resisters and transistors, found Ok. I tried different tricks to go to calibration or setup mode but couldnt. Now days due to COVID-19 the scanners are short in the market and price goes very high > 250USD
hoping to get some help from you, as company is not responding
Not sure what you went with but just sample your area and not each pixel. I'm thinking like 6 points to 9 points.
how would you calibrate that thermometer? Are the offset values stored in that eprom?
It would have a calibration routine where, on the production line, each device is pointed at a black body that is adjusted to be as close as they can reasonably get to an exact temperature (probably something like 0 degrees C, 20 degrees C, etc up to 100 degrees C) which the on-board processor can then do a basic linear regression to help offset differences in individual sensors, assuming they are roughly linear. It's possible they do many more temperatures than that, or fewer that are further between. And yeah, the exact voltage at each temperature would be stored in that EEPROM. Without knowing exactly how they do their calibration routine, it'd be pretty difficult to re-calibrate yourself. Theoretically you could probably get a more accurate device if you did have the time and materials to make very precise temperature targets.
afaik it's voltage peak to peak for AC just like Veff for (1/sqr(2))*U
does the lens increase the range of the sensor
Sort of, but there's a limit with fresnel lenses. If you want to really "zoom in" with a thermal camera you'd need some pretty expensive silicon optics.
Is there a way to calibrate it?
Basically, no.
You sound like Gabe from The Office (Tv Series) :)
i think you can build one using a webcam changed to inflared (use negative photo filter) and a software that mesure distance and the contrast and density of the pixels (make a simple mathematic formule to keep this information always available) calibrate all things and you have a webcam that read distance and temperature with incredible accurancy✾☄
It's not that easy, unfortunately. Refer to the wikipedia article on "Infrared Spectroscopy"
How, exactly, does a negative photo filter (not even sure what you mean by that) turn a visible light sensor into an infrared sensor?
I didn't get your point.
@orbiter8 If you like this kind of stuff, search for 'eevblog' on here. His videos are nice and technical too.
Patents are stupid... those ideas are such... a kid could think of them