DIY-Thermocam V2: A do-it-yourself thermal imager

jvandive:
Wow!
This is exactly what I have been looking for! I was looking for something to look for thermal leaks around my doors and windows. I was wondering when the enclosue and PCB might be ready for purchase? I can't wait to build this project! Thank You very much!

Jvandive

As I am currently out of money (student life is expensive :slight_smile: ), I cant test the new version of the case (with modified servo mount) and the newly created PCB at the moment :frowning:
If you don't want to wait 1-2 months until i become "rich", you have 2 posibilities:
I give you the order links for both and a instruction on how to put everything together (should be a lot more easier than the current building process on the website), but there may be the possible risk that the PCB is not working correct (the case should be fine anyway).
Second, you send me the material and shipping costs in advantage. I order the components for you, put everything together here and try if it works. If I find any mistakes, i will solve them and if that is not possible immediately, I order a second revision of the pcb/case and pay it on my own. At the end, I sent the whole package to your home adress.

Maxbot,

I have already ordered most of the parts needed to build the Thermocam. I also wanted the fun of building it myself so, I guess just send me the links for the PCB and the case and I will order them. I will take the risk on the PCB not being correct.

Thank You in advance!

Jvandive

jvandive:
Maxbot,

I have already ordered most of the parts needed to build the Thermocam. I also wanted the fun of building it myself so, I guess just send me the links for the PCB and the case and I will order them. I will take the risk on the PCB not being correct.

Thank You in advance!

Jvandive

pm sent. i hope everything works :slight_smile:

Maxbot

The new software is a big step up on the old stuff. Well done. I noticed that when there is a hot object in the view the cool objects are lost in the image because the degree difference is not enough to show up on the final display. May I suggest you put a button on the final display to re-compute and re-display the data "logarithmically" to compress it and show up the finer details if so desired. It also might be a good idea to have a custom scan resolution window so you can enter something like 5X64 for a quick scan to check for termites in the wall. That's what I'm trying to do with your scanner in my house. Every two or three years they attack it.
Anyway, great work.
Cheers
Stephen

inmyshed:
May I suggest you put a button on the final display to re-compute and re-display the data "logarithmically" to compress it and show up the finer details if so desired.

I currently have no time to get into this. If you have a mathematical approach how to compress the data logarithmically, I could implement it fast.

inmyshed:
It also might be a good idea to have a custom scan resolution window so you can enter something like 5X64 for a quick scan to check for termites in the wall. That's what I'm trying to do with your scanner in my house. Every two or three years they attack it.

Good idea, I will include that in the next version.

I came across your project and find it to be very interesting. I'm wondering how your getting the actual image from a basic web camera without adding the necessary lense to see the the correct wavelengths? Is what I'm seeing in the images a simulated thermal image through software?

code101:
I came across your project and find it to be very interesting. I'm wondering how your getting the actual image from a basic web camera without adding the necessary lense to see the the correct wavelengths? Is what I'm seeing in the images a simulated thermal image through software?

No, this would not be possible. There is no filter for seeing thermal infrared, you need an ir sensor or thermopile for this.
I only added the webcam so that you can see an optical image of the scanned area. The MLX90614 infrared sensor then moves above that area like a scanner and measures the temperatures at many points. The computer software processes those data and creates a "thermal image" by transforming temperature values into specific colours.
That means you can not get a live stream of images, each takes time for the one single sensor to get across the area.
I hope you now understand what my project is about :wink:

I am having a problem.

I built the thermocam by following the instruction from cheap-thermocam.com

Thermal sensor works, and it measures the temp nicely.

The problem is that, when i try to take the picture, it is keep saying "wrong com-port" blahblah...

I tried to install the driver again, used another program, and when i click left, right, up, and down, servos move accordingly...

but when i try to start running the program, i'm keep having the same problem.

ANy suggestions?

Does anyone know how to connect this thermocam using xbee? (instead of using usb cable)

juntom10:
Does anyone know how to connect this thermocam using xbee? (instead of using usb cable)

If you can use xbee, then you should not ask this. It is probably connected the same way as with USB.
Did I misunderstand the question?

Cheers,
Kari

Hi all -

I found this project a couple of weeks ago on the web and I though it was very cool and totally worthwhile replicating. So, I'm at the point where I'm trying to program the configuration sketch into the Arduino Uno. BTW, I have the Uno R3 which is all that Sparkfun is selling these days. Anyway, the configuration sketch is loaded and when I open up the serial monitor window I'll see the loaded program generating message to "enter any key" to the start the process of re-configuring the EEPROM. So, I'll enter a letter into the input box and click send then I'll see the message "erasing EEPROM" or something like that and there it will hang. I've let it sit in this state for about 15 minutes and nothing else happens. How long should it take to run the configuration sketch once installed into the Uno? If anyone can help I would be most appreciative.

Best regards,

Chris.

Hi,

Residents of countries which have a comma as "official" decimal separator and for whom
Cheap-Thermocam or ThermalCamera software does not work: start the software specifying
a more sensible locale i.e.:

java -Duser.country=EN -Duser.language=en -jar ThermalCamera.jar

Or fix the default system locale settings.

Cheers,

Seppo

I may have answered my own question. Seems I'm jumping the gun a bit with the software part. I thought the configuration sketch was for the Arduino but on closer inspections of the code comments, it's actually for the thermal sensor eeprom which is not connected yet. Good deal. I'll connect everything together then re-run the config sketch and see if that works.

Best regards...

Chris.

Very nice project ! 8)

Hello,
This is probably one of the best projects I have ever come across, and I almost have all of the necessary parts to begin my build. I had one question,

Did you ever consider the technique of "stacking" a number of scans? Even though the data is not video, it is pixel centric. The technique increases the signal-to-noise ratio and increases the dynamic range , by pixel. Here, a 'pixel' is a data measure point in temperature, but it is related to a specific area that has been measured, just like a pixel of RGB would be in video. The technique uses averaging and summing, so that the image can be , say put into a higher resolution , much like in photoshop, you take a low resolution image and convert it from 72dpi to 300dpi, you get much more definition. This works by using a stacking algorithm from the data in one image only, but true stacking lets you create a bitmap of higher resolution, and then fill in the data using from 2 to up to a hundred similar scans run through the stacking algorithm. The concept is explained below in this astronomy website that covers the technology fairly well:

http://keithwiley.com/astroPhotography/imageStacking.shtml

When I get my system built , I planned on trying it, however I wanted to run it by you to see if it made technical sense in this context.

Thanks so much for what you did. Hope to hear from you soon.

CharlyV

This sounds really interessting.
I printed that page out and will read it through as soon as I have time :slight_smile:

charlyv:
Hello,
This is probably one of the best projects I have ever come across, and I almost have all of the necessary parts to begin my build. I had one question,

Did you ever consider the technique of "stacking" a number of scans? Even though the data is not video, it is pixel centric. The technique increases the signal-to-noise ratio and increases the dynamic range , by pixel. Here, a 'pixel' is a data measure point in temperature, but it is related to a specific area that has been measured, just like a pixel of RGB would be in video. The technique uses averaging and summing, so that the image can be , say put into a higher resolution , much like in photoshop, you take a low resolution image and convert it from 72dpi to 300dpi, you get much more definition. This works by using a stacking algorithm from the data in one image only, but true stacking lets you create a bitmap of higher resolution, and then fill in the data using from 2 to up to a hundred similar scans run through the stacking algorithm. The concept is explained below in this astronomy website that covers the technology fairly well:

How Image Stacking Works

When I get my system built , I planned on trying it, however I wanted to run it by you to see if it made technical sense in this context.

Thanks so much for what you did. Hope to hear from you soon.

CharlyV

Any chance this sensor is the exact same one?

I don't want to plunk the money down on something from China that I'm not sure is going to be right when I can get a BCI version for the same price here in the US.
My main concern is I don't know if I can program the EEPROM of the BCI the same way as the DCI. If I can, I would rather get the BCI since it'll get here faster from a reliable source.

Update: All parts for the new version are here - MLX90620 sensor is coming on Monday.

Trying to create the menu for the display now with touch implementation and sd store function :slight_smile:

For those following this project, does anyone want to do a group buy on the new sensor? It's about $65/sensor if you buy one but if we can get 10 people in the CONUS to go in on a group buy, we could save about $8-9 in the end.

maxbot:
Update: All parts for the new version are here - MLX90620 sensor is coming on Monday.

Trying to create the menu for the display now with touch implementation and sd store function :slight_smile:
Can you share with us the part list of the new version?Nice image

Thanks a lot for your great camera.