Actual resolution of Flir One V2

[TopLoser]

The new Flir One as it stands is perfect for all my needs. Easy, convenient, decent resolution and now a -20 to +120 degree temp range. And all for £160 ish + VAT.

Everybody I show it to wants one for that price, a new 'must have' gadget for those that have never seen one before.

[tomas123]

there is a flir one v2 (android) review with some nice raw images
http://www.slashgear.com/flir-one-thermal-camera-for-android-hands-on-30391243/

s sample:

 
the code:

Code: [Select]

//extract RAW and resize 300%
exiftool -b -RawThermalImage 1.jpg > v2.png
convert  -define png:swap-bytes=on v2.png -auto-level -filter point -resize 960x 960x720.png

//get palette
exiftool 1.jpg -b -Palette > pal.raw
convert -size 224X1 -depth 8 YCbCr:pal.raw -separate -swap 1,2 -set colorspace YCbCr -combine -colorspace RGB pal.png
convert 960x720.png pal.png -clut 960x720color.png

//  160x120 (x6 -> 960x720)
convert -size 6x6 xc:none -stroke gray -strokewidth 0  -draw "line 0,0 0,5"  -draw "line 0,0 5,0"  6x6.png
convert -size 960x720 tile:6x6.png -transparent white grid6.png
convert 960x720color.png grid6.png -flatten overlay-160x120color.png
convert overlay-160x120color.png -crop 200x150+540+500 -quality 100 crop-160x120color.jpg

//  120x90 (x8 -> 960x720)
convert -size 8x8 xc:none -stroke gray -strokewidth 0  -draw "line 0,0 0,7"  -draw "line 0,0 7,0"  8x8.png
convert -size 960x720 tile:8x8.png -transparent white grid8.png
convert 960x720color.png grid8.png -flatten overlay-120x90color.png
convert overlay-120x90color.png -crop 200x150+540+500 -quality 100 crop-120x90color.jpg

and the result:

160x120


120x90


To my mind the resampling algorithmus of this image is really better.
I cant see the sensor resolution on diagonal lines (typical steps) and noise artefacts.
It's possible that this pre-production (android) model is not software (firmware) crippled.

[Fraser]

The android software is still under development as we know. The iOS software would have had to have been submitted to Apple for approval some time ago and needs to be re-approved with each update.

Could it just be that we are seeing a new version of processing algorithms that have yet to be released for the iOS model ?  There is also the possibility that the 2nd Gen ONE for iOS is running software based on the old 80x60 iOS FLIR ONE version. The Android software will likely be all new and potentially created by a different team of coders ?

The banding on the prototypes images is unusual but may well be a software issue that still requires 'treatment'.

If the images produced by the pre-release Android version are already better than those that we have seen from the iOS I am hopeful that we may see great things from the 2nd Gen One Android in July 

Aurora

[tomas123]

the above image converted with level stretching (min. temperature spreading) shows interesting artefacts
 

Code: [Select]

convert 960x720.png -level 20000,50000 1.png
The image quality is great!!

[tomas123]

the raw image is ovesharpen

The banding on the prototypes images is unusual but may well be a software issue that still requires 'treatment'.
Aurora

I think, it's a bad NUC (Non Uniformity Correction) function

from user manuel

The camera has a mechanical shutter inside that periodically activates
and allows the thermal camera to do a calibration or image refresh.
When the shutter activates, the image will freeze for a brief period.
The purpose of the shutter is to allow the camera to provide an
improved image. The FLIR ONE performs the tuning operation
automatically on a periodic basis.

... I am hopeful that we may see great things from the 2nd Gen One Android in July 
Aurora

like Flir One v2 connected to a Linux PC (i.e. a raspberry pi webcam) 

[Trax]

like Flir One v2 connected to a Linux PC (i.e. a raspberry pi webcam) 

does that really work?

[tomas123]

sounds like a minor hack with the Flir One SDK and an USB adapter (OTG?)
http://developer.flir.com/flir-one-software-development-kit/

[Trax]

Ok my phone is most likely not flir compatible even after installing a 5.1 ROM it does not allow me to download the flir one apk from the pay store :/
could someone send it to me for testing if it installs.

With the 5.1 rom it however allows to install the seek thermal software.

So if there is no simple way to operate the flir from a windows or Linux PS/Table I will have to go for the seek damn it.

[Fraser]

Why do you have to go for the SEEK ?...... I think you will regret such a decision.

I bought a used phone cheaply off of ebay for my SEEK experiments and I will be using it for my V2 FLIR ONE. It cost me £30 for an as new Motorols MOTO G. It was sold cheaply as it is a US Verizon phone that does not work as a phone in the UK. I also bought a MOTO X for £20 because it had a faulty speaker. That is no good however as it is not 'On the Go' capable.

I often see various mobile phones sold with faulty phone parts such as 'will not read SIM' or speaker broken etc. Take a look fro a decent condition one that suits your needs.Do not pay too much though    The Wi-Fi and Bluetooth connectivity is normally still fully operational in these phones..

Aurora

[Trax]

>Why do you have to go for the SEEK

Cause its more likely to work with my phone, and if not there are sources in an other thread here some-ware to with a windows app that would allow me to use it on any of my x86 windows tablets.

I haven't looked on ebay for cheep compatible phones,
but if you look on the flir website the tested phones are only very few top of the line units. Of cause it will very likely work also with sufficiently many older cheep units but without knowing which once its like playing roulette.

Or do you think that all thats needed is USB OTG and a kernel >= 4.3

My LG has OTG support only with a custom kernel so possibly google thinks than my phone does not have OTG at all and thats why it denyes the installation even with a 5.x ROM.

I think the best would be if someone would find a quick way to use the Flir One on a normal PC/tablet with windows/linux.

Trax

[Muxr]

You could also get an iPod touch. Not sure if it's a cheaper alternative, but you would be able to order the iOS version which is a few weeks wait I believe.

[Trax]

You could also get an iPod touch. Not sure if it's a cheaper alternative, but you would be able to order the iOS version which is a few weeks wait I believe.

My aversion towards apple is to large for that. I already get pustules when I have to use MacOS in a VM.
If I would have to use an actual crapple device my tow nails would probably role up or something similar.


But on a serious note I really like x86 hardware and something that runs on almost any android has good chances to be runnable under windows/linux on a PC as happened with the seek.
So I definitely will go for something that is android compatible.
I don't mind the waiting I have enough other work to take my mind of the waiting.

[Muxr]

Heh I gave in to Apple many years ago . Ran Linux prior to that for many years. The older I get the more I find I'd rather spend my time tinkering with my projects then keeping my computer and phones running the latest hack or whatnot. Apple devices are restricted, but at least they work well out of the box.

I understand though

[dadler]

Off-topic but locally on topic: Same with me. I exclusively used Linux (and sometimes Windows) from the mid-nineties until about 2007. A customer complained that a web application I had developed did not work in "Safari".

So I begrudgibgly went to the Apple Store and bought a Mac Mini to test the app against. And the rest is history: desktop Unix that generally "just works"--and doesn't require constant hand-compilation of the kernel or kernel modules.

Been using Mac when I can ever since. I'm torn on iOS vs Android though. I own devices of both platforms, but ordered the iOS flavor of the Flir One v2.

[tom66]

With an added lens I found the Seek to be more than usable for my electronics work.

I'm likely to sell my Seek I acquired from Aurora soon and move to a Flir One, but I'm still pretty happy with the Seek. I think with some better software it would be a good alternative to the Flir. The sensor may not be as good but for general purpose electronics work (finding shorts, determining hotspots, etc.) it works exactly as one would require.

[tomas123]

a minor hint:
Flir double the (extrapolated) image size of the embedded raw values from Flir One version 1 to 2.

Code: [Select]

$ exiftool -RAWTHERMAL* FLIR_ONE*
======== FLIR_ONE.jpg
Raw Thermal Image Width         : 120
Raw Thermal Image Height        : 160
Raw Thermal Image Type          : PNG
Raw Thermal Image               : (Binary data 19971 bytes, use -b option to extract)
======== FLIR_ONEV2-ANDROID.jpg
Raw Thermal Image Width         : 320
Raw Thermal Image Height        : 240
Raw Thermal Image Type          : PNG
Raw Thermal Image               : (Binary data 52476 bytes, use -b option to extract)
======== FLIR_ONEV2-IOS.jpg
Raw Thermal Image Width         : 240
Raw Thermal Image Height        : 320
Raw Thermal Image Type          : PNG
Raw Thermal Image               : (Binary data 80870 bytes, use -b option to extract)
Source:
Flir One V2 Android http://www.slashgear.com/flir-one-thermal-camera-for-android-hands-on-30391243/
Flir One V2 IOS https://www.eevblog.com/forum/testgear/actual-resolution-of-flir-one-v2/msg701478/#msg701478

[encryptededdy]

Looks like FLIR have released another product using the 160x120 lepton core - the FLIR K2 (firefighting camera).

According to this press release: http://investors.flir.com/releasedetail.cfm?ReleaseID=916817

Powered by FLIR's revolutionary Lepton(R) camera core, the compact K2 delivers detailed thermal images at 160 x 120 resolution

On the specs page here: http://www.flir.com.hk/fire/display/?id=68967

It tells us the unit is 9Hz and the lens is f/1.1

Temperature range is –20°C to +150°C (–4°F to +302°F) and 0°C to +500°C (+32°F to +932°F). Presumably the first range is "native" to the sensor, and it's what we see on the FLIR C2 or raw lepton, whereas the 2nd range is using a ND filter.

[wischi]

How are the FLIR ONEv2 Macro capabilities for taking photos of PCBs with the stock lens?

Did someone already try the ONEv2 with an additional ZnSe lens?

Any advice on what to get for taking thermal images of PCBs? SEEK + ZnSe lens, SEEK xr or ONEv2 + ZnSe lens?

Best,
wischi

[Muxr]

With the standard lens it's usable but it's a bit par for the course.

The board pictured is 87x63mm.

[Pantheron]

Hello,

i found some documents around the new FLIR Lepton Core (3), with its improved resolution but smaller Pixelsize etc..
Also an Databrief, appnote where are the 80x60 and the 160x120 compared.

So everything is the same in the core, except the resolution and this is also affects the vospi.. Now you have the frames (80x60) at 120 Hz.. they send 4 frames at 80x60 so you have to "assemble" it in 1 frame with 160x120..
But everything is described in the documents.

http://www.mako.co.kr/#!---/cyai

So its nice to read, its just plug in to get 160x120 where actually is 80x60 !!

Regards
Pantheron

[mikeselectricstuff]

Nice info - links to actual PDFs
Lepton 2 vs. 3 differences
http://media.wix.com/ugd/53cdb6_9b698564a2e84af6a955078615b3a9da.pdf
Lepton 3 Datasheet
http://media.wix.com/ugd/53cdb6_5191be73d1c943d78d2e1a095cb7f3b8.pdf


Interstingly the new one has some NV memory for storing configurations - I suspect this is the extra chip seen on Toploser's teardown

[Fraser]

@Pantheron

Very interesting reading. Thank you 

So FLIR have also gone for a 12um pixel size. I still wonder if the patent dispute with SEEK Thermal related to work FLIR was also doing on 12um microbolomter technology and techniques.

As expected FLIR are using some nice image processing in this LEPTON 3 Core. I like the detail that they have provided on FFC and ALC operation. This core also has dynamic dead pixel mapping  (not just at time of OEM calibration).

I may be ahead of myself here, but FLIR have a real beauty to market in the form of this new LEPTON 3 core    If it does not encourage 3rd party usage of thermal imaging in an array of products then nothing will.

Aurora

[mikeselectricstuff]

Wonder when we'll see them in Digikey

[tomas123]

Flir doesn't puplish informations, how to convert the "Radiometry Enabled" RAW14 stream to temperature [°C / °F].

I think, this is a big benefit of a calibrated Flir One V2 (with a real camera).

the calibration datas from a Flir One V2 image

Code: [Select]

>exiftool -Planck* -Atmospheric* -Offset*  FlirOneV2\FLIR_ONEV2-ANDROID.jpg
Planck R1                       : 11208.23
Planck B                        : 1311.1
Planck F                        : 1
Planck O                        : 174
Planck R2                       : 0.01068296
Atmospheric Temperature         : 20.0 C
Atmospheric Trans Alpha 1       : 0.006569
Atmospheric Trans Alpha 2       : 0.012620
Atmospheric Trans Beta 1        : -0.002276
Atmospheric Trans Beta 2        : -0.006670
Atmospheric Trans X             : 1.900000
Offset X                        : +3
Offset Y                        : -7
Maybe it's simple to hack the USB Port of the Android version to get this stream.


two sample images from different FlirOneV2 with the thermal calibration parameters:
https://www.eevblog.com/forum/testgear/actual-resolution-of-flir-one-v2/msg701478/#msg701478
https://www.eevblog.com/forum/testgear/actual-resolution-of-flir-one-v2/msg706841/#msg706841

There is also a calibration value for the y-offset between thermal to real camera!

Code: [Select]

>exiftool -Planck* -Offset* *.jpg
======== 1.jpg
Planck R1                       : 15314.136
Planck B                        : 1404.6
Planck F                        : 1
Planck O                        : -978
Planck R2                       : 0.011565135
Offset X                        : -1
Offset Y                        : +34
======== 2.JPG
Planck R1                       : 15633.827
Planck B                        : 1410.8
Planck F                        : 1
Planck O                        : -1655
Planck R2                       : 0.012125017
Offset X                        : +1
Offset Y                        : -6

[wischi]

Ok, one more kinda unrelated question: Do you guys feel the +120°C*vmax is to low for PCB-analysis?

Best,
wischi