Yet another cheap thermal imager incoming.. Seek Thermal

[eneuro]

Every few seconds image gets "nicer".
This are two frames at 00:02 and the difference is huge:

Nicer does not mean more accurate, but it can be simply blured as well
In the case of static images (not moving object sin the scene) I could add option in software to do moving average preset amount frames back, so one could get oversampled thermal imaginery and when something is not moveing we should get nice thermal improoved images, while there is no problem to mount camera and point to area of interest.

[eneuro]

Wire just published review: http://www.wired.com/2014/10/seek-thermal-infrared-camera-iphone-android/

For me interesting is this  Thermal+ mode 

"But wait! There is one other mode – Thermal+. In this mode, the app uses both the IR camera and the iPhone’s built in camera. This way you can switch back and forth between visible and IR images."

Hopefully, those two cameras are on the same side, but why this app only shows them side by side-they do not try to apply edges to thermal image?
Flir's MSX patent issues or simply this crappy phone visual light camera has very diffrent focus not compatible with Seek Thermal?

It could be interesting to see difference in optics of those two cameras, however I do not plan to buy iPhone no way, so more details about Seek Thermal optics could be very interesting to know, to... find proper visual light camera to be able apply those edges from visual to thermal 

[sacherjj]

Just got shipping notice. Left Boise yesterday. Has next week delivery listed.   Order number in 600s.

[Fraser]

@sacherjj,

Good news 

Aurora

[callipso]

Mike, would you happen to know your order number?

[mikeselectricstuff]

Mike, would you happen to know your order number?

69x - it has arrived in the US, the acknowledgement email arrived afterwards!

[callipso]

Cue pinouts and protocol specs...  O0

Also hihi, 69

[iRad]

Just got a message from Seek saying that they are very sorry they have not shipped my (Apple) Seek Thermal order in the time originally expected. Adding that it longer to get their apps in the app stores than anticipated, and that the delay created a significant backlog of orders. Further stating that after a brief component shortage temporarily halted production, they are as of this week ramping up production and once again shipping Android cameras.

Seek says that they will likewise begin shipping Apple cameras once the app has been approved in iTunes, which they hope will be very soon... "in the next few weeks."

[miguelvp]

I got the has shipped email at 2:12 pm CDT and the order update apologizing at 2:35pm, since I was cheap I wont get it until Wednesday.

[marshallh]

Mine will be here Monday

[waldo]

Every few seconds image gets "nicer".
This are two frames at 00:02 and the difference is huge:

That is due to the periodic recalibration.  You can hear the shutter click each time the image pauses.  On my unit, the fixed pattern noise slowly increases for a few seconds and then suddenly goes away (mostly) each time it clicks.  I have never used a thermal camera before, but I not not expect such a quick drifting of the image quality.  As I recall, Mike did not recalibrate in his little DIY camera using a Lepton, so the Seek sensor appears to be different from a Lepton in that regard.

[Fraser]

Regarding NUC events..... extant technology that requires an NUC event tends to operate teh NUC shutter quite often while teh imager chip and surrounding electronics come up to nominal temperature. If you look at a microbolometer using another thermal camera you will see that it produces heat and runs at around 30C. While it warms up teh NUC is needed to compensate for teh large delta T that is occurring. Once at normal operating temperature, teh NUC events should become less frequent and teh nominal setting in many cameras for routine NUC is 120 seconds. On Industrial cameras you can often change teh fixed interval of even switch off teh NUC for videa work etc. My NEC Avio cameras offer this menu option as standard.

Without an NUC event the Inductrial cameras take some time to suffer the mottling effect that is a sign of temperature drift in teh Pixels. I have had a stable image for more than 5 minutes BUT thise cameras have TEC temperature stabilised microbolometers. The E4 drift rate is much rater due to teh lack of any TEC but the software does try to compensate based on the chassis and microbolometer die temperature sensor readings.

The SEEK appears to suffer noise issues and this may well be due to a lack of processing in the associted software. Much of FLIR noise reduction technology is software based. The SEEK microbolometer is using the smallest pixel size that I have seen at 12uM. There is a down side to small pixel size. Less surface area means less sensitivity in much teh same way as CCD pixels suffer from lowe sensitivity as they become smaller. With CCD chips teh technology has compensated fro this. Small pixel size and subsequent lower mass may also add to temperature uniformity issues across the array.

The bottom line is that teh SEEK is breaking new ground and using a new smaller FPA for teh microbolometer. It is also using a mobile phone host and reliant upon software for imaga eenhancement. IMHO it would seem unreasonable to expect the SEEK to have an equivalent image quality to one of my Industrial cameras taht have desicated SBC's for the OS and image improvement algorithms. I would expect its image quality to improve as the software develops over time however. I am surprised that, with the higher resolution, it does not provide far superior images to the FLIR One however. MSX does add visyual detail but even teh thermal images of the SEEK do seem to suffer from severe noise artifacts. This will need to be addressed.

When all is said and done, if you want a very cheap thermal imaging capability for your phone, the SEEK seems to fit teh bill. If you want decent image quality and a neat self contained unit, please consider the FLIR E4. There MAY be good news for E4 Firmware 2.1 and 2.3 owners soon   

Aurora (floating around in the Atlantic!)

[mikeselectricstuff]

Tracking shows (one of) mine has arrived in the country 

[Legit-Design]

Without an NUC event the Industrial cameras take some time to suffer the mottling effect that is a sign of temperature drift in the Pixels. I have had a stable image for more than 5 minutes BUT these cameras have TEC temperature stabilized microbolometers. The E4 drift rate is much rater due to the lack of any TEC but the software does try to compensate based on the chassis and microbolometer die temperature sensor readings.

Small pixel size and subsequent lower mass may also add to temperature uniformity issues across the array.

So would modifying it with a TEC (Peltier) possibly make it better? Just to cool it down enough so it doesn't start condensation but still keep it at optimum temperature. I've seen some cool modifications for astro imaging cameras using peltiers, just a small cold finger will do wonders.

[iRad]

GREAT NEWS FOR IOS APPLE USERS... THE APP IS NOW IN THE APP STORE!!!      

[eneuro]

THE APP IS NOW IN THE APP STORE!!!

APP with closed source code is not needed, but real hardware to hack 

[mikeselectricstuff]

Without an NUC event the Industrial cameras take some time to suffer the mottling effect that is a sign of temperature drift in the Pixels. I have had a stable image for more than 5 minutes BUT these cameras have TEC temperature stabilized microbolometers. The E4 drift rate is much rater due to the lack of any TEC but the software does try to compensate based on the chassis and microbolometer die temperature sensor readings.

Small pixel size and subsequent lower mass may also add to temperature uniformity issues across the array.

So would modifying it with a TEC (Peltier) possibly make it better? Just to cool it down enough so it doesn't start condensation but still keep it at optimum temperature. I've seen some cool modifications for astro imaging cameras using peltiers, just a small cold finger will do wonders.

I don't know too much about the physics of the sensors,  but I'm sure that if cooling would help, manufacturers would be doing it, as the cost of a peltier is small compared to the sensor.
I suspect one issue is that you don't want to create thermal gradients across the sensor, which would be hard to avoid.
Sensors are packaged in a vacuum to avoid convection losses, so it wouldn't be hard to mount the die on a peltier inside a sealed package to avoid condensation issues.     

[iRad]

THE APP IS NOW IN THE APP STORE!!!

APP with closed source code is not needed, but real hardware to hack   

APP is needed, because availability of the iOS APP means that real hardware will now ship.

[mikeselectricstuff]

Just got hold of mine. Having issues possibly due to old phone -in the meantime can someone download the latest app  for me (not available outside US) - looks like it was updated Oct 14th

[sacherjj]

Just got hold of mine. Having issues possibly due to old phone -in the meantime can someone download the latest app  for me (not available outside US) - looks like it was updated Oct 14th

Check your PMs...

[frenky]

Just got hold of mine. Having issues possibly due to old phone -in the meantime can someone download the latest app  for me (not available outside US) - looks like it was updated Oct 14th

Can't wait for the teardown & reverse engineering video...

[mikeselectricstuff]

Just got hold of mine. Having issues possibly due to old phone -in the meantime can someone download the latest app  for me (not available outside US) - looks like it was updated Oct 14th

Check your PMs...

Thanks!

[cynfab]

I got mine on Fri. about 10 minutes before the shipping e-mail arrived.
Plugged it into my Galaxy S4 Mini and..... nothing. fiddled around  for a while turning on & off USB debugging. Re-plugging, re-installing, registering..... still nothing. Was it a dud???
Well not quite. Plugging it into a linux desktop lsusb produced: Bus 002 Device 006: ID 289d:0010
Sideloading the app onto my Nexus 7 didn't produce any results either. So off goes an e-mail to their tech support.
As it was Fri afternoon, I wasn't expecting any response. I wasn't disappointed, so I had all weekend to stew.
Not being able to leave it alone, I dug out my Radxa Rock, and after much fussing got a bootable SD card with KitKat 4.4.2 running. The app appeared on the Play Store, so I installed it and plugged the camera into the OTG port... nothing, just like the S4  Mini.
But wait, the Rock has 2 USB host ports and plugging the camera into either one, (or the USB hub I have my mouse/kbd in) and the app asked if it could use the USB port (I said yes) then the app said it was updating the camera firmware ;>) then the app gave me a really nice black screen ;>(.
But wait, all the controls were un-greyed and by pressing the camera and shutter I got a real thermal snapshot. I was not able to record a video though.

It will be interesting to see what their tech support says about all this..
And their app rev 1.2.1 crashes a lot and sometimes you must clear the user data to stop it from crashing on startup.

The attached image is just as it came off the Radxa Rock, exif and all. Notice that it is 832 x 624 pixels.

   ...ken...

[mikeselectricstuff]

There have been reports of flakiness on other platforms, in particular Samsung S3, and a message on their facebook page indicates that they have found some issues and will be issuing an update soon.
 

[marshallh]

Empty space where Lightning<>USB bridge interface sits.
Case is aluminum, pried apart not too much fuss.

Someone mail me a crappy anrdoidr phone and I'll get usb logs