[Steve] sent in a tip to show us his Thermography scanner. Constructed from an Arduino, two servos, a thermal sensor and a little bit of code, it is fairly simple. The results aren’t groundbreaking. You can see his examples are fairly low resolution and took about 30 seconds to capture. It isn’t bad for a quick project though. The source code is available on his site.
nice proof of concept,what about a rotating thermal sensor,kinda POV but reverse,anyone dare?
Sweet! he followed on with my idea by re-purposing his original distance measuring sensor with a thermal sensor – (see the comments) http://hackedgadgets.com/2009/04/02/3d-arduino-scanner/
I’ve been wanting to try this for a while buy the £50 price tag for the 8+1 thermopile has meant it’s taken a back burner. http://www.robot-electronics.co.uk/acatalog/Thermal_Array_Sensor.html
I can’t open the link
What no beetle?
Yay crappy web hosting! :P I hope it’s back up soon, I want to see it…
Ohhhhhhhhhhhh Noes, there goes Tokyo, GO GO ARDUINO!!!
Seriously though, nice hack. How long till one of the “Arduino sucks” trolls joins the discussion?
arduino sucks… just kidding (coleman) I dont know why people have to hate on the arduino so much people are making cool things with it that other wise would be not made period and it is a pretty good platform
@”why people have to hate on the arduino so much people are making cool things with it that other wise would be not made period and it is a pretty good platform”
because it lowered cool to: rotating 2 servos and reading AD
Cool hack!
Thats exactly what i was thinking of for an IMU. get a horizon reference with this method. Sky is cold freezing 0c to very cold around -30 or worse, land is less cold to even warm during the summer. That should make an epic clear horizon image.
Hello,
the Link above has a small mistake, its working but not giving the right layout…
http://www.designer2k2.at/index.php?option=com_content&view=article&id=30:thermoscanner&catid=13:arduino&Itemid=40
that should work :)
awesome job Steve keep up the good work.
thermal cameras are so dam expensive and you dont need extreme details all the time.
I don’t know why (therian) has to be so jealous lol.
Nice job!! I’m definately going to try this.
Maybe you could get rid of some of that periodic interference by reading multiple values from the sensor and averaging them, might be induced by electric fields. If the sensor really takes 40ms for 9px that makes about 4.5ms/px, which is well in range for 50/60Hz line voltage.
the TPA81 has only 1°C resolution, and the values can jump +-1°C while nothing changes. Thats why the pictures are so noisy…
im working on a averaging, if you take 5 samples the result gets already a lot smoother.
I hope that with going to 10 or more samples i could get below 1°C resolution, making it capable to detect defect insulation on our house :)
Nice hack Steve, could care less about the arduino since in this case using just the AVR would seriously complicate this project.
Sidenote – Anyone else notice this sensor costs over a hundred bucks.
Hey, why is the comment I posted 12+ hours ago still waiting moderation?
Andrew, you are right ;)
but there is no other thing around what could do the job, i couldnt get a single thermopile with optics, and the cheap IR thermometers have a way to big FOV.
If somebody can give me a hint to a cheap thermopile with a very narrow FOV (smaller as 5×5°) i would take it right away :)
Well I found the company that makes the sensor used in the TPA81. They also make a 4×4 and since the TPA81 has little more than the array and a PIC, this might be a fun project.
http://optoelectronics.perkinelmer.com/Catalog/Category.aspx?CategoryName=TPLM+%2F+TPAM+Series
@TheBadWolf
The scan is probably too slow for your idea of a rotating disk with a row of sensors (that’s how I understand your idea), the disk would then have to rotate very slowly and you’d get the same results as this, except to have a row you’d have to buy a ton of sensors and all interface them and scan those simultaneously.
Oh and disk would have the issue of the space near the center being condensed compared to the edge due to the rotation.
Although you could just make a sliding bar with a row of sensors I guess, but it would still be expensive and a hassle to interface and read probably
What about bouncing the light off a hexagonal (or octagonal, or whatever) mirror scanner, like you’d find in a laser printer?
A piece of metal with a slit in it would probably be enough to reduce the FOV, but the signal strength (and the SNR as well) will likely go through the floor…
My point being that it’s easier to move a balanced mirror quickly than to move a (relatively) unbalanced thermopile array…
Couple two mirrors together (and maybe a few lenses), and it should be possible to get some form of 2D scanner rigged up. Scan speed would depend on how fast you could read from the thermopile array chip, though…
alternate approach:- you can get very small PIR sensors, just make a row of them and scan that sideways using a stepper. then scan the array vertically, and build up a composite picture as a row of scans.
same idea as a video head…
I don’t get why people are using a overbloated microcontroller for things that could’ve been done with way less parts. if he built this w/o any microcontroller it’d be cool. But so … not very impressive.
@conundrum: no you can’t. PIR sensors sense heat movement across the sensor, not absolute temperature.
@xl – it’s because they’re not as cool as you.
There I go again, feeding the trolls.