Author Topic: Yet another cheap thermal imager incoming.. Seek Thermal  (Read 750945 times)

0 Members and 2 Guests are viewing this topic.

Offline frenky

  • Supporter
  • ****
  • Posts: 969
  • Country: si
    • Frenki.net
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2200 on: November 21, 2015, 12:10:39 pm »
Update...

As said before pixel values stabilise after a half an hour or so. So when pointed to isothermal plane the images come out pretty nice and noiseless.

But when pointed to different temperature than the one of isothemal plane, each pixel gain is different. So the resulting images are very noisy.

This is the image of the room corner (aprox 2*C difference):


Taken with my app:


Taken with seeks android app:


Now that I understand Seek sensor much better I'm also able to get best possible image from Jadew's app:



As you can see. Jadew's app is still better because it deals with different gain of the pixels in a way that noise is not visible...

So where to go from here... I love thermal cameras but I'm done with trying to improve seek.

I'll rather spend my free time finishing a few other project which will bring me  some €€€ so that I can buy ThemalExpert from i3system when it comes out in December.
When I get it I'll put all my effort in developing best custom app for it...
« Last Edit: November 21, 2015, 12:13:32 pm by frenky »
 

Offline Nik

  • Contributor
  • Posts: 9
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2201 on: November 21, 2015, 08:02:18 pm »
But when pointed to different temperature than the one of isothemal plane, each pixel gain is different. So the resulting images are very noisy.
Is this noise the same pixel to pixel on each frame?
 

Offline frenky

  • Supporter
  • ****
  • Posts: 969
  • Country: si
    • Frenki.net
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2202 on: November 21, 2015, 08:13:24 pm »
Yes because its not a real noise it's just the fact that each pixel has different gain.

For example if the average value of pixels at 20*C is 7500. When you point Seek at plane with 22*C most of the pixels will go to 7700 but approx 10% of the pixels will jump to 8000. All values are approximate.



Sent from my LG-D855 using Tapatalk

 

Offline arsenix

  • Contributor
  • Posts: 24
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2203 on: November 23, 2015, 03:33:42 am »
The variation in gain is not something that is unique to Seek. It just matters how well the manufacturer has calibrated the device. No inexpensive device is going to be particularly well calibrated in this department. Even higher priced uncooled devices are rarely exhaustively calibrated unless they are extremely high dollar. It is particularly challenging to calibrate since the gain (for each pixel) also varies as the device temperature changes. I am working with the Flir Lepton and it displays similar behaviour (although their built in NUC seems better at filtering it out).

I have found that calibrating with an isothermal plane at a few different temperatures allows for fairly consistent results. Those papers I linked earlier in the thread describe some more exhaustive techniques.

This is an image I took with the Lepton (raw sensor) where I did a two point calibration to convert each pixel value to an absolute temperature. Much of the "noise" goes away when you do this to compensate for offset and gain variations per pixel.

 

Offline frenky

  • Supporter
  • ****
  • Posts: 969
  • Country: si
    • Frenki.net
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2204 on: November 23, 2015, 09:00:34 am »
Tnx for the info. :-+ I might give it a shot when I'll have time...
 

Offline tomas123

  • Frequent Contributor
  • **
  • Posts: 832
  • Country: de
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2205 on: November 23, 2015, 09:22:04 am »
@arsenix:
The lepton RAW values are linear to Radiance with an offset.
The result is not exact linear to temperature, because of fourth power of temperature. See curve in excel screenshot :)
I wrote here something about the calculation of Flir RAW values to temperature and posted two excel sheets:
https://www.eevblog.com/forum/testgear/yet-another-cheap-thermal-imager-incoming/?action=dlattach;attach=176884
https://www.eevblog.com/forum/testgear/flir-e4-thermal-imaging-camera-teardown/?action=dlattach;attach=168737


This is an image I took with the Lepton (raw sensor) where I did a two point calibration to convert each pixel value to an absolute temperature. Much of the "noise" goes away when you do this to compensate for offset and gain variations per pixel.


that the response of "A/D Counts" are linear to Radiance  with an offset.
one day later and no reaction  :-//

for better understanding I made a excel sheet

the Flir E4 has a spectral range from 7.5–13 um (the peak 10.6 um) http://support.flir.com/DsDownload/Assets/63901-0101_en_50.pdf
 
I wrote a table for the planck law at 10600 nm wave length over the temperature range
There you can see, that the spectral radiance in [W / m^2] (per nm wave length) is not exact linear



source: http://www.ces.karlsruhe.de/culm/culm/culm2/th_physik/simulation/planck.xls
« Last Edit: November 23, 2015, 09:26:04 am by tomas123 »
 

Offline arsenix

  • Contributor
  • Posts: 24
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2206 on: November 24, 2015, 03:18:53 am »
That is good stuff! I have not incorporated that into what I am doing yet (but I intend to). I am mostly focused on small temp ranges (20C swing total), so doing a few points and linear interpolating seems to be accurate enough so far (and definitely removes most of the gain artifacts).

Actually from your chart it looks like the spectral radiance is much more linear than the response of the microbolometer pixels to radiant energy, at least for what I have seen in the research and from measurements on the device. Based on my measurements, Lepton RAW values are certainly not linear to radiance with an offset. Gain varies by huge amounts from pixel to pixel. I don't have my data handy at the moment, but I believe variations were 5-10C without compensating for per pixel gain variation.

I actually didn't do a two point calibration it was more of a three point. I took a flat plane at near ambient temp (for the offset) and then did one at 10C below and 10C above that temp. It interpolates linearly at those points. Not exhaustive... but close enough to do what I want for now. What I found with this technique does match the trends in your spectral radiance graph though. Lower temperatures had less counts/degree, higher temps had higher counts/degree. When I sit down again with this next week I'll try to incorporate your spectral radiance curve. Maybe this will linearize the gain per pixel over temperature (although I believe it will still vary from pixel to pixel).
 

Offline tomas123

  • Frequent Contributor
  • **
  • Posts: 832
  • Country: de
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2207 on: November 24, 2015, 11:00:39 am »
Lepton RAW values are certainly not linear to radiance with an offset.

I'm sure, that also the Lepton Sensor (like all Flir Sensors around 10um wavelength) gives an (14 Bit) RAW signal which is linear to radiance with an offset (if you activate the feature "radiometry enabled"). 
In minor temperature range this is nearly linear.

There is a nice hint in this Flir documents:
http://flir.custhelp.com/ci/fattach/get/1667/
on page 2 they say, that the response of "A/D Counts" are linear to Radiance with a offset.

Currently Flir doesn't public more infos about the Leptons radiometric:
http://www.flir.com/cvs/cores/knowledgebase/index.cfm?CFTREEITEMKEY=914&view=67485

... but I know that the Flir One G1 (Lepton 1) and Flir G2 (Lepton 3) saves 14 Bit RAW images, which are linear to radiance.

The formula [Kelvin]=f([14 Bit RAW]):
I posted here the derivation: http://u88.n24.queensu.ca/exiftool/forum/index.php/topic,4898.msg23944.html#msg23944
Quote
http://en.wikipedia.org/wiki/Planck%27s_law

spectral radiance I of a body per unit wavelength

I = (2*h*c^2/L^5)/(e^(h*c/(k*L*T)-1)

Planck’s Constant h = 6.626068x10-34 joule sec
Boltzman’s Constant k =1.38066x10-23 joule deg-1
Speed of light in vacuum c =2.997925x10+8 m/s
T = object temperature in Kelvins
L = wavelength in m (lampda)
I = radiance in Joules/m^3/sec/steradian
e Euler Number

with substitute
R1 = 2*h*c^2/L^5
B = h*c/(k*L)

it goes to
I = R / (e^(B/T)-1)

solve it to Temperature
T= B / ln(R1/I +1)
with ln => Natural logarithm

In Flir cameras is radiance "I" linear to  Flir A/D Counts!
we can write
I = R2*(S+O)
with
S = 14 Bit A/D Counts
R2 = constant factor
O = offset

goes to
T = B / ln(R1/(R2(S+O))+1)

with R=R1/R2:
T = B / ln(R/(S+O))+1)  :-+


but  simple use the Flir document from attachment.
Code: [Select]
Temperature (in Kelvin) = B / log(R / (S - O) + F),
where S is the 14-bit digital signal value.
log(x) is the base-e logarithm of the x parameter.

Set F=1 (used for all Leptons) and calculate B,R and O with a temperature measurement over >=3 known points.
see document:
The offset O is around 1500
Planck B value range 1300 - 1600.

If you haven't a black body, then you must consider the emission from ambient source  8)



PS: more infos you find in the FLIR GEV Demo Source Code (1.8.0)
http://80.77.70.144/SwDownload/app/RssSWDownload.aspx?ID=143
PvSimpleUISampleDlg.cpp
Code: [Select]
USHORT PvSimpleUISampleDlg::TauToSignal(float C)
{
    USHORT s;
float K = C + 273.15f;
    s = (USHORT)(m_R /(exp(m_B/K) - m_F) + m_O);
    return s;
}

float PvSimpleUISampleDlg::TauToTemp(USHORT sig)
{
    float T;
    double objSig;
    objSig = powToObjSig((double)sig);
    T = (float)(m_B / log(m_R /(objSig - m_O) + m_F));
    return T;
}
« Last Edit: November 24, 2015, 03:05:31 pm by tomas123 »
 

Offline arsenix

  • Contributor
  • Posts: 24
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2208 on: November 26, 2015, 01:59:37 am »
I will try that. I did not attempt to plot or otherwise examine the per pixel variations in detail yet so it is possible that the data will fit to that formula. I should have saved the reference images I took with my flat field temperature reference, but I can take additional shots.

Unfortunately we can't enable radiometry mode on the Lepton (yet). Information is thin but it seems that radiometry mode on this device just changes the offset relative to device temperature. With radiometry mode enabled, offset is independent of device temp, with it disabled offset is dependent on the device temperature so that half output is approximately the current device temperature. I believe (and hope) that this means that calibration can be performed even with radiometry disabled as long as device temperature is kept relatively constant (or calibration is performed at multiple temperatures of interest).

Appreciate the data dump!
 

Offline tomas123

  • Frequent Contributor
  • **
  • Posts: 832
  • Country: de
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2209 on: November 26, 2015, 10:43:18 pm »
Unfortunately we can't enable radiometry mode on the Lepton (yet).

I didn't know that the SDK has no radiometry switch.
I only read the lepton data sheet, and there is the radiometry mode explained (see"8.4.2 Radiometry Enabled").

The Flir One G1/G2 with Lepton 1/3 use this radiometry mode. I think, it's only a marketing strategy.

PS: I found this
http://www.flir.com/cvs/cores/knowledgebase/index.cfm?view=67336
Quote
Can I make use of Lepton's Radiometry Module to make absolute temperature measurements?
Availability of Radiometry commands is restricted to Original Equipment Manufacturer (OEM) customers for Lepton, and requires that an active non-disclosure agreement (NDA) be in place with FLIR.
 
The OEM quantity threshold is 1,000 units for Lepton.
>:D
« Last Edit: November 26, 2015, 10:48:16 pm by tomas123 »
 

Offline arsenix

  • Contributor
  • Posts: 24
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2210 on: November 27, 2015, 02:24:33 am »
Yeah exactly... the radiometry mode command is a secret unless you are a high volume customer!

Do you think Flir One uses it? I thought it probably just used the shutter (which has a temp sensor on it) to do temperature offset correction. If Flir One uses the mode it should be fairly easy to sniff I2c to figure out the command.

Based on other information though Flir has claimed that only the Lepton 3.5 (not released quite yet) will support the radiometry mode fully. I believe another question in that FAQ says this.

Anyways... further discussion on this should probably be on a Flir thread :)
 

Offline tomas123

  • Frequent Contributor
  • **
  • Posts: 832
  • Country: de
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2211 on: November 27, 2015, 08:27:23 am »
Anyways... further discussion on this should probably be on a Flir thread :)

You are right, but we have multiple Flir One threads but no Lepton thread here.  :-\

Offline in.Vitro

  • Contributor
  • Posts: 46
  • Country: de
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2212 on: November 27, 2015, 04:05:11 pm »
One question: Is SeeK dead? :S
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 8730
  • Country: gb
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2213 on: November 27, 2015, 05:42:35 pm »
Seek Thermal are offering a 25% discount on their range at the moment. If that doesn't bring in some sales then they have a problem.

Fraser
 


Offline brobbuilder

  • Contributor
  • Posts: 15
  • Country: as
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2215 on: December 03, 2015, 10:19:41 pm »
Can only extract so much useful data from a tiny shitty sensor. Better to buy old FLIR i3 on EBay and hack large microbolometer to give 320x240. i3 is like DSLR compared to Gameboy camera Seek.
 

Offline miguelvp

  • Super Contributor
  • ***
  • Posts: 5549
  • Country: us
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2216 on: December 03, 2015, 10:31:36 pm »
yawn  :=\
 

Offline encryptededdy

  • Frequent Contributor
  • **
  • Posts: 358
  • Country: nz
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2217 on: December 04, 2015, 01:15:30 am »
Can only extract so much useful data from a tiny shitty sensor. Better to buy old FLIR i3 on EBay and hack large microbolometer to give 320x240. i3 is like DSLR compared to Gameboy camera Seek.
Yes it has a 320x240 microbolometer, but I think you can only hack the i3 to 240x240, which is equal to it's screen resolution.
 

Offline brobbuilder

  • Contributor
  • Posts: 15
  • Country: as
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2218 on: December 04, 2015, 01:55:51 am »
ix (not ix0) series screen is only 140px across. Same as i7 resolution. I'm thinking you should take out the bolometer and just hack that (or the FPGA). Forget the screen.
 

Offline Trax

  • Regular Contributor
  • *
  • Posts: 119
  • Country: at
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2219 on: December 24, 2015, 10:10:47 pm »
What do you think about the new Seek Thermal Reveal, is it worth getting one?

A Flir one is for me not an option as my smartphone apparently don't support USB host

EDIT: or how about this one:; http://www.peaktech.de/produktdetails/kategorie/waermebildkamera/produkt/peaktech-5605.html

What I'm looking for is a thermal camera for below 500€ incl tax, that does not require a smart phone
« Last Edit: December 24, 2015, 10:21:02 pm by Trax »
 

Offline Trax

  • Regular Contributor
  • *
  • Posts: 119
  • Country: at
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2220 on: December 27, 2015, 08:19:24 am »
no one an opinion on the topic?
 

Offline -jeffB

  • Regular Contributor
  • *
  • Posts: 114
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2221 on: December 27, 2015, 03:03:52 pm »
We'd need to know more about your intended use.

The SeeK Reveal IS a thermal camera, and it IS under 500€.

It will show you things that are hot and cold.

It won't show small temperature differences very well.

It won't give very accurate temperature measurements -- at least, my first-generation SeeK dongle doesn't.

What do you want to do with it?

For that matter, how long do you expect to keep your smartphone before upgrading to one that does support USB host?
 

Offline Trax

  • Regular Contributor
  • *
  • Posts: 119
  • Country: at
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2222 on: December 27, 2015, 05:30:18 pm »
> It will show you things that are hot and cold.
I figured that much ;)

> It won't show small temperature differences very well.
What is small?

I found a video here: where the host compares some < 1500€ cameras including the seek.
At the end there is a compaction list he say that that booth the seek and the flir one can resolve temp differences of about 0.4 °C

> It won't give very accurate temperature measurements -- at least, my first-generation SeeK dongle doesn't.
Yea the guy noticed that as well.
Supposedly the flir one head an error of +2% at 50°C and +1.6% at 120 °C
while the seek head an error of -8.7% at 50°C and -5.3% at 120°C as well as -4.9% at 350°C

The reveal had an error of 0 at 120°C but the other cases ware comparable so I guess that was a coincidence.

I think for me the absolute value is not so important if I want to measure a spot exactly I can use something else, sooting differences, hotspots, overall temperature distribution is more interesting.

> What do you want to do with it?
Once you got a hammer every problem looks like a nail ;)

No but seriously:
1) I think what I'll be doing first is check my breaker box if something is running hot, may be check other parts of my electrical installation
2) I surly will look for thermal leaks in my house, a.k.a. where to fix some isolation etc.
3) I may possibly want to use it when playing around with my 3d printer since i added a heated enclosure I would like to see the temp distribution around the extruder. I noticed sometimes that the filament gets stuck as apparently its getting to soft for the motor to push it. So at this point some improved thermal barrier will be needed.
4) Looking on PCB's if/where they are running hot, I can imagine that may be quite useful like for finding failed tantal caps, etc...
5) Finding other problems i can solve with a thermal cam.

> For that matter, how long do you expect to keep your smartphone before upgrading to one that does support USB host?
One or two years at least, its not even two years old now.
And I really don't see any reason to switch it, it has a replaceable battery, and still more CPU power and Memory that I need right now.


Also Its a rather expensive tool, and if its smartphoen based its not guaranteed to work forever once the manufacture goes out of busyness (no app updates). Well one option would be to get a cheep compatible phoen for this purpose only install the app make a flash backup and keep it offlien forever :D
 

Offline -jeffB

  • Regular Contributor
  • *
  • Posts: 114
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2223 on: December 28, 2015, 01:57:48 pm »
> What do you want to do with it?
Once you got a hammer every problem looks like a nail ;)

Now you're speaking my language.  ;D

No but seriously:
1) I think what I'll be doing first is check my breaker box if something is running hot, may be check other parts of my electrical installation

SeeK should work for that. Temperature differences > 5 C or so should be easy to see.

2) I surly will look for thermal leaks in my house, a.k.a. where to fix some isolation etc.

SeeK might work for that. In my experience, insulation problems usually show as a small temp difference. SeeK will show leaks around windows and doors easily, but one or two degrees across a wide stretch of wall will be very hard to distinguish from noise. I've been spoiled by a modified FLIR E4, which very clearly shows even a fraction of a degree's difference. The SeeK's images look terrible when the temperature range across the scene is narrow.

3) I may possibly want to use it when playing around with my 3d printer since i added a heated enclosure I would like to see the temp distribution around the extruder. I noticed sometimes that the filament gets stuck as apparently its getting to soft for the motor to push it. So at this point some improved thermal barrier will be needed.

Might work for this, if the temperature differences you want to see are several degrees or more. The images will look better because of the contrast between the heated parts and the unheated surroundings; the resulting wide temperature span will minimize the appearance of noise.

4) Looking on PCB's if/where they are running hot, I can imagine that may be quite useful like for finding failed tantal caps, etc...

Possibly. For this use, it would be a good idea to get a device with adjustable focus, and/or to buy/build a close-up lens attachment.

5) Finding other problems i can solve with a thermal cam.

There's never any end to those. :) I found a spot on my ceiling a fraction of a degree colder than the rest; it's obvious with the E4, difficult but possible to spot with the SeeK. It apparently shows a small leak in an upstairs toilet drain. So, the camera can show you things you wish you hadn't seen.

> For that matter, how long do you expect to keep your smartphone before upgrading to one that does support USB host?
One or two years at least, its not even two years old now.
And I really don't see any reason to switch it, it has a replaceable battery, and still more CPU power and Memory that I need right now.


Also Its a rather expensive tool, and if its smartphoen based its not guaranteed to work forever once the manufacture goes out of busyness (no app updates). Well one option would be to get a cheep compatible phoen for this purpose only install the app make a flash backup and keep it offlien forever :D

Sounds like keeping your current phone is a good idea. But a phone-attached camera is more flexible for capturing and sharing images and video. When I checked earlier in the year, there were several cheap (US$50-100) tablets that could run a SeeK dongle. They aren't as convenient as a well-designed all-in-one unit, or something plugged into a hand-sized phone, but they do work.
 

Offline joe-c

  • Frequent Contributor
  • **
  • Posts: 323
  • Country: de
    • Joe-c.de
Re: Yet another cheap thermal imager incoming.. Seek Thermal
« Reply #2224 on: January 16, 2016, 11:00:28 pm »
My program is still not finished yet, but the image quality is better than earlier versions.
Maybe interesting for somebody  :-/O
FLIR E4 hardware mods 
Freeware Thermal Analysis Software: ThermoVision_Joe-C
Freeware Remote Software for some FLIR Cameras: Ex Tools
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf