Flir E4 Thermal imaging camera teardown

[KK]

@KK

I have both the SEEK and the E4 and can definitely recommend the E4 over the current SEEK offering. You will not be disappointed with an E4. I know the SEEK is much cheaper but sometimes you have to spend a little more to get decent image quality especially now that the firmware 2.3 can be upgraded to 320x240 pixels.

Aurora

Thanks for the info. The attraction to the Seek was its small size and attachability to my iPhone/iPad. The Flir-1 doesn't fit by iPhone 6+ so that's out. I have always liked the E4 I just needed the higher resolution and now that is a done deal.

[KK]

@KK

I have both the SEEK and the E4 and can definitely recommend the E4 over the current SEEK offering. You will not be disappointed with an E4. I know the SEEK is much cheaper but sometimes you have to spend a little more to get decent image quality especially now that the firmware 2.3 can be upgraded to 320x240 pixels.

Aurora

For the moment for $800 more than Seek one gets with Flir E4 $1000 only thermal blobs enhanced by higher resolution visual camera  as official product
amazon: FLIR E4: Compact Thermal Imaging Camera with 80 x 60 IR Resolution and MSX
This is officially available upgrade supported by Flir or only hacking cracking crying, etc,  to get this magic 320x240 thermal scene?
If only hacked version than you have to pray with each Flir software upgrade if you will be able hack this resolution in next versions too, else this thing will be able only show thermal blobs at oryginal 80x60 resolution 
Seek with gradient issues corrected and own open source software which will enable something similar  to Flirs  MSX easy and at much higher resolution can be Flir killer soon, while no strange menu hacking needed and everyone can write his own application and add functions that will never be available in Flir software.
For the moment without any development tools it is possible get realtime thermal imaginery from Seek Thermal Android connected to PC Linux/Windows or Android tablet -only one need to make adjustments to those gradients in the corners, but it is clear that Seek will also improve hardware soon and while even now it is available to use this dongle at Windows and Linux systems with own software without any hacking oryginal software it is much more interesting option if someone wants more than only this what comes with Flir soft...

I see your point, but remember nobody HAS to update the firmware. If some future firmware comes out, you don't NEED to update if you don't want to. Any resolution hacks your already have will always be there if you don't update.

[gemrough]

I see your point, but remember nobody HAS to update the firmware. If some future firmware comes out, you don't NEED to update if you don't want to. Any resolution hacks your already have will always be there if you don't update.

Exactly. And if you shop around you should be able to get an E4 (or E4B) for $800 or less. Yes out the box the E4 is only enhanced by MSX... and the SEEK is a competitor (just based on price, it kicks the E4's butt)

However if you enable the full 320x240 via software modification the E4 blows the SEEK out of the water!

[eneuro]

However if you enable the full 320x240 via software modification the E4 blows the SEEK out of the water!

Via hardware modyfication and connecting 5 Seek Thermal dongles for $1000 you get two times bigger thermal pixels array - excluded 2.1k hexagon patent pixels per Seek device, so about 30k Seek Thermal real available pixels vs 76.8k at this higher E4 resolution, but still crappy 640x480 visual camera in E4    which is nothing to available 5M or even 8M in modern Full HD visual cameras available.
It will depend on application-Seek it is much smaller and ready to connect to existing small PCB Linux boards and make functional thermal objects deection system accelerated by Nvidia Cudo supported by OpenCV 

[Bud]

Ok so i get a firmware upgrade for my E4 and what do i -really- get? I am not getting box measurements, i amnot getting manual scale, all i get maybe some minor bug fixes best case, and worst case more locked firmware, so who cares about firmware upgrades?!

[Fraser]

I am sorry to have to say it but anyone who believes that the SEEK microbolometer is in the same class as a FLIR 320x240 microbolometer is kidding themselves. We have to remember that the SEEK exists for one reason only....to make thermal imaging affordable for the masses. It is a heavily compromised solution and necessarily so in order to meet the very low retail price. For the money the SEEK is interesting to me as a new technology but its image noise content and gradient issues makes it an early days prototype IMHO and not a polished product. I certainly do not see a SEEK strapped to a tablet or PC running a 3rd party software as a very ergonomic device for many. It needs its issues in both hardware and APP to be sorted and sorted effectively. It must be remembered that FLIR are past masters in thermal imaging and the processing of the images before output to screen or computer.

Can I recommend the SEEK product at this time ? Sadly No. It breaks new ground but if you are serious about using a decent thermal camera, for non industrial  use, you should seriously consider the E4 upgraded to 320x240 and accept that it will be without a warranty if FLIR detect the hack upon return for repair. The very good news is that FLIR equipment is normally very reliable unless abused.

Maybe the SEEK Mk2 (if it ever happens) will do better ? I would expect a higher retail price though.

Aurora

[earzone]

switch to a different computer... I would recommend Win XP if you have access to one.

YEAH!! thats it 
this was the solution 
thanks @ all for your support!

greets

[mikeselectricstuff]

Even an un-hacked E4 still beats the seek hands-down - MSX very effectively augments the low native resolution apart from in close-up situations.
Even then the much better temp range/sensitivity gives a "nicer" picture than the seek.

[Emoticon]

Installed the menu hack without issue on a 2.3 1.2L. Even this small upgrade feels like the day I got the camera and installed the hack all over again! My wife even complains that we needs a new rule that no thermal cameras will be on the dinner table . The difference between the rainbow color pallet and rainbow HC is like night and day! Studs inside walls show up much better than they did before. I also notice if you hold the camera so that your skin is the only thing the camera can see, then while holding it there put the temp range on lock (increases contrast in the 90-100F range), set black hot and pull the camera back you can now see the veins in your arm very well in black hot specifically! Thank you so much guys, the utility you've added to this device is nothing short of amazing! It's such a letdown that FLIR isn't producing the best product they can, but instead intentionally crippling their products.

[eneuro]

Even an un-hacked E4 still beats the seek hands-down - MSX very effectively augments the low native resolution apart from in close-up situations.

Since there is only global setting for emissivity even in "perfect" Flir E4 with a few different emissivity objects in the scene high precision will be useless since one can't specify emissivity on scene object level, so output temperature will be affected by global emissivity setting, so Seek dongle with automatic object recognition and applying emissivity to predefined objects can easy outperform overpriced Flir$ especially when connected to Nvidia Cudo accelerated machines including multicore PC and tablets with Nvidia graphics cards 

BTW: Gradient issue tested in a way just by putting this Seek dongle very close to surface by putting on table reversed is bad idea while sensor can see reflected heat from himself, so it was clear mistake which was made in this famous @mikeselectricstuff  Seek thermal gradient video .
I understand, this might be made for Flir friends, but simply this not the use case this Seek thing was designed to-cover lens input with plate without any minimal distance and enjoy gradients. No serious scientific experiments were presented by @Aurora and @mikeselectricstuff how this gradient changes with distance from lens and surface angle

[miguelvp]

@eneuro,

The gradient issue in the Seek is there, there is no doubt about it and it develops even if pointing at the floor at a distance.

The two people you point out are probably the most experienced with thermal imaging devices in this forum.

[eneuro]

@eneuro,

The gradient issue in the Seek is there, there is no doubt about it and it develops even if pointing at the floor at a distance.

The two people you point out are probably the most experienced with thermal imaging devices in this forum.

@miguelvp Are you taking about oryginal app? How do you know that it was not already solved by.... me in my custom software?  i have no time to wait for official Seek development tools as well for Flir, while for <<$1000 custmers will get complete solution not only Flir E4 with degraded thermal resolution to 80x60 and visual 640x480 
I'm not using oryginal Seek Android app while it also has fixed global emissivity as it is passed as not matrix but one float parameter in this Seek Andoid app code in this native call to libSeekware.so installed with Seek app on Android

libSeekware.so: ELF 32-bit LSB shared object, ARM, version 1 (SYSV), dynamically linked, not stripped

Code: [Select]

SeekwareNativeLib.ThermographyCalculate(bytebuffer, bytebuffer1, floatbuffer, 32448, 32448, 0, 0, 208, 156, [b]emissivity[/b], minMaxTemperatureBuffer, bytebuffer2);
So, simply Flir E? with  fixed global emissivity  doesn't fit my needs, but it looks like Seek can do easy  for me at decent resolution  of course in my own software

[miguelvp]

Yes I'm talking about the app. Also I'm not sure what you are talking about solving anything since the code I released doesn't show that gradient and I did share how I did it with source code a long time ago and I still have to see any algorithm from you.
https://www.eevblog.com/forum/testgear/yet-another-cheap-thermal-imager-incoming/msg546880/#msg546880

Even the lastest drop that uses a different approach doesn't show the gradient either, and I did share the code as well.
https://www.eevblog.com/forum/testgear/yet-another-cheap-thermal-imager-incoming/msg547638/#msg547638

I'm also aware about the methods available in libSeekware.so since I did point to you (just yesterday) that the decompiled code you dismissed the poster as just an apk, unzipped.
https://www.eevblog.com/forum/testgear/yet-another-cheap-thermal-imager-incoming/msg556179/#msg556179

Maybe you are also going to claim what row 207 does as well?

In fact the praise should go to marshall and sgstairs for communicating with the Seek anyways.

Now, BS aside, and I've been biting my tongue till now, You can't claim the E4 is not superior than the Seek because you have neither of those cameras and you come out in an insulting way towards people that are more knowledgeable than you in this matters (mike and aurora). Might be a language barrier but I doubt it because your constant eye rolling towards others.

And you (and myself) should stay away from the E4 thread. They have a wonderful sensor that works stand alone and this thread should be left stand alone as well, wanna discuss it further let's take it over PM.

[SCR]

I am planning to purchase an E4 via Amazon to apply the resolution hack but am a bit overwhelmed by the amount of postings and comments on this topic, so excuse me for lazily taking the shortcut: can the E4's which are currently sold be modified or are there any (hw/sw) versions which cannot be modified (e.g. newer version than 2.3?)?

Cheers!

[eneuro]

Yes I'm talking about the app. Also I'm not sure what you are talking about solving anything since the code I released doesn't show that gradient and I did share how I did it with source code a long time ago and I still have to see any algorithm from you.

In next posts like this Since the ring is not centered it could be due to the lens placement and the software assumes the ring is supposed to be centered it looks like you have a lot of to do to do it without manual tweaking for every Seek Thermal camera 
Just investigating nonlinear gradients in Flir E? 320x240 in similar spots too-need more 16bit gray PNGs from Flir equipment , before they make thermal blobs enhanced by MSX contours

[gemrough]

What the hell does any of this have to do with the E4 hack or how to help unlock more features in 2.3.0??? Oh, it doesn't!! Your just being an annoying troll!!!

Yes I'm talking about the app. Also I'm not sure what you are talking about solving anything since the code I released doesn't show that gradient and I did share how I did it with source code a long time ago and I still have to see any algorithm from you.

In next posts like this Since the ring is not centered it could be due to the lens placement and the software assumes the ring is supposed to be centered it looks like you have a lot of to do to do it without manual tweaking for every Seek Thermal camera 
Just investigating nonlinear gradients in Flir E? 320x240 in similar spots too-need more 16bit gray PNGs from Flir equipment , before they make thermal blobs enhanced by MSX contours

[gemrough]

I am planning to purchase an E4 via Amazon to apply the resolution hack but am a bit overwhelmed by the amount of postings and comments on this topic, so excuse me for lazily taking the shortcut: can the E4's which are currently sold be modified or are there any (hw/sw) versions which cannot be modified (e.g. newer version than 2.3?)?

Cheers!

Amazon is shipping 2.3.0 and you can do a partial hack to it. You can unlock full resolution and also add several palettes to the menu. Some features are still locked at this time.

[Bud]

We are not aware of versions newer than 2.3.0
I guess you better hurry up before one is out

 

[Bud]

Seek dongle with automatic object recognition and applying emissivity to predefined objects can easy outperform overpriced Flir$ especially when connected to Nvidia Cudo accelerated machines including multicore PC and tablets with Nvidia graphics cards

My friend you must be kidding about pricing comparison when you saying what SEEK has to be connected to. Secondly, good luck drugging your CUDA (not Cudo) PC around on your next house thermal inspection trip.

[samf]

Thank you all for helpful insights and contribution. Special thanks to "gemrough" who replied my post.

Today I successfully upgraded my E4 (1.1L, 1.22.0) to E8 without much problem, except that when I did first time I lost my MSX feature using itools1, but later when I re-tried, it came with new CRC number and it was fine. I did both resolution and Menu hack.

The "Special edition" logo appeared first time and after that it never appeared and USB mode still enabled after restart, also I see new option "Hidden Menu".

Can I leave it in USB mode? If not how to change it to original status.
What should be the normal operating mode for the camera after the hack? RNDS/MSD/UVC ?

Thank you all.

[eneuro]

Secondly, good luck drugging your CUDA (not Cudo) PC around on your next house thermal inspection trip.

Common we live in XXI century and ARM Linux is everywhere now  even on small PCBs
http://beagleboard.org/black

Flir is not needed to see and record in realtime thermal image -a few lines of code with Seek dongle can do what we want, not what Flir wants 

BTW: Today "trolls" use at least C/C++/java languages and also know IBM AS/400 iSeries ILE RPG  and of course  Linux very well

[Bud]

Photos of unconnected boards do not count, my friend. Noone questions your programming skills but programming is not everything. It's a long jorney from a concept to manufacturing. Please do return when you have a _finished_ product and we will be happy to talk .

[tomas123]

Flir is not needed to see and record in realtime thermal image -a few lines of code with Seek dongle can do what we want, not what Flir wants 

this is not true
see my post
https://www.eevblog.com/forum/testgear/flir-e4-thermal-imaging-camera-teardown/msg377140/#msg377140

the flir raw video format is *.seq and can simple decoded with exiftool
http://u88.n24.queensu.ca/exiftool/forum/index.php/topic,5279.msg25567.html#msg25567

[pomonabill221]

What the hell does any of this have to do with the E4 hack or how to help unlock more features in 2.3.0??? Oh, it doesn't!! Your just being an annoying troll!!!

Yes I'm talking about the app. Also I'm not sure what you are talking about solving anything since the code I released doesn't show that gradient and I did share how I did it with source code a long time ago and I still have to see any algorithm from you.

In next posts like this Since the ring is not centered it could be due to the lens placement and the software assumes the ring is supposed to be centered it looks like you have a lot of to do to do it without manual tweaking for every Seek Thermal camera 
Just investigating nonlinear gradients in Flir E? 320x240 in similar spots too-need more 16bit gray PNGs from Flir equipment , before they make thermal blobs enhanced by MSX contours

I cannot agree MORE!!!  Why the Hel* is this jerk trying to yank everyone' chain about the Flir being so inferior and the $eek being SO much better??

IT IS NOT!!!
So GO AWAY!!!!!  if you continue to be a pain in the A$$!!!  that you are!!!
I am getting TIRED of reading your assinine posts!!! that have NOTHING to do with the Flir teardown/(hack) that makes these cameras MUCH better than the $eek will EVER BE!!!

[SCR]

Amazon is shipping 2.3.0 and you can do a partial hack to it. You can unlock full resolution and also add several palettes to the menu. Some features are still locked at this time.

Thank you, will try to get one this week and then report back