Author Topic: Any modern 320x240 Thermal Imager comparable to the Fluke Ti32 in Image Quality?  (Read 1306 times)

0 Members and 1 Guest are viewing this topic.

Offline Faith

  • Supporter
  • ****
  • Posts: 154
  • Country: sg
Hi everyone!~ So it's been a really long time since I've last played around with a thermal imager and I've been pondering buying one again in the future.

Here's my dilemma though: I used to own a Fluke Ti32 back when it was still a current model and those sold for like 8,000US$ back then but I was really lucky to get one BNIB for half that thanks to someone clearing inventory.

I eventually sold it for a profit though I did really enjoy my time with it.

So I've been looking around to see what's changed in the ten or so years since I let go of my Fluke Ti32, and while it's pretty easy to find a thermal imager for around a grand with the same resolution these days, I noticed that it's hard to compare the image quality.

A lot of the cheaper thermal imagers seem to be super noisy, and even though they claim 320x240 resolution they're nowhere near as crisp as the images I used to take with the decade old Fluke Ti32.

Not to mention the Fluke Ti32's sensitivity is still pretty impressive by today's standards (≤ 0.045 degrees C and 45 mK) and the hi-lo span went as low as 2.5°C.

I really don't want to go down in resolution (nor sensitivity), and I doubt I'll have the budget for anything higher (though 640x480 would be awesome haha).

Any recommendations as to what I should be looking for?

I'd also love to mention that the firmware on the Fluke Ti32 was a dream. It was a super smooth and fluid unit to work with. Fluke really do make awesome thermal imagers and I did always prefer them over FLIR.
<3 ~Faith~
 

Offline Spirit532

  • Frequent Contributor
  • **
  • Posts: 429
  • Country: by
    • My website
The FLIR Exx series(not Ex!) has quite a good image quality, the T4xx series(I have one, a bit biased) is marginally better than even the Exx, and the newer E53, E75, E95, T5xx, T6xx, all have a 640x480 sensor that is crippled in some way or another, but it's substantially better than the competition.
On the cheaper side, the 384x288 Therm-App is pretty good.
Not going to suggest anything lower, since you're asking for image quality that would compare to an older generation Ti32.
 

Offline Faith

  • Supporter
  • ****
  • Posts: 154
  • Country: sg
Thanks for your reply!~

Yeah the FLIR E53 looks worthwhile even though the resolution is lower than I'd like (the E75 is definitely out of budget).

Problem with the FLIR Exx Series is that it's kinda in the price range where adoption among hobbyist is low so it'll probably be quite a while before someone finally picks one up to see if it can be hacked up haha.
<3 ~Faith~
 

Offline Spirit532

  • Frequent Contributor
  • **
  • Posts: 429
  • Country: by
    • My website
The E53 may be hackable to 640x480, but nobody knows since it really isn't out in the wild.
Get one into the hands of someone on the forum who is willing to work on hacking :)
 
The following users thanked this post: uski

Offline calel

  • Regular Contributor
  • *
  • Posts: 87
  • Country: ch
the newer E53, E75, E95, T5xx, T6xx, all have a 640x480 sensor that is crippled in some way or another
ok how theh do you know that? is that even true or just a rumour?
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9909
  • Country: gb
Calel,

FLIR operate the most efficient manufacturing process that they can achieve when producing thermal imaging cameras at it increases the profit margin. With the Ex and Original Exx series they used the same hardware and BoM for all the cameras so that only one hardware platform needed to be manufactured.. there were optional modules like Wi-Fi but they could easily be fitted after main production. The Ex5 series appeared on the market and presented some unusual resolutions that are not available in standard microbolometers. These immediately suggest that a larger microbolomter has had its output manipulated to provide a non standard resolution. There are two ways to reduce a microbolometers resultion, as presented to the user..... you can “Window” the microbolomter and select how much of the arrest is read out, or how much of the read out field is used. This is effectively a form of digital zoom and it effects the FOV of the produced scene. This is a crude resolution hobbling method that I have seen on L3 Thermal Eye based cameras. It is simple to apply, but also can be simple to reverse  ;) The other option is manipulation of the image in the fully digital domain to down sample/scale a higher resolution array to behave like a lower resolution. This option permits the generation of non standard resolutions but care is needed with the down sampling/scaling to avoid significant image degradation for non simple scaling ratios. The advantage of this resolution reduction method is that it offers different resolutions whilst maintaining the same. FOV using a common lens specification across a range of cameras and thermal scene resolutions.

When looking at the specifications for the Ex5 series we see an unusual resolution is offered by one model yet the FOV remains the same for a lens of the same specification as the other cameras. This would suggest that the thermal resolution is a result of downsampling or scaling and not windowing. Knowing what we do about how FLIR operate, it is not unreasonable to believe that the Ex5 series use the FLIR 640 x 480 pixel microbolometer across the whole series and just reduce its resolution through down sampling/scaling to meet a specific cameras specification. I personally have not had confirmation that this is the case from FLIR so cannot confirm that belief as fact but the evidence Does support it.

So what are the other possible scenarios ? Well FLIR make the TAU2 core in different resolutions using the QVGA, QVGA+ and VGA microbolometers. Could they just be using one of the TAU2 microbolometers to suit a particular cameras resolution ? The answer is yes, it is possible that they could do this, but unconfirmed and not supported by the datasheet specification. The highest resolution E95 camera offers 464 x 348 which is a non standard resolution. That tells me that the E95 does indeed use a VGA microbolometer that is down-sampled to this unusual resolution. Can the full resolution be unlocked ? That very much depends upon how a FLIR have protected the down-sampling method and whether the camera is calibrated at its microbolometers native resolution or at the lower, down-sampled, resolution. I suspect full native VGA resolution Calibration would be simpler for FLIR though as down sampling/scaling can create anomalies !

Evidence from the data sheet of a VGA microbolometer and down sampling/scaling :

The Datasheet states that the microbolometer in all models uses 17um pixels.
The interchangeable lenses used across the series provide the same FOV in all models.
A QVGA 17um based Microbolometer would provide approx half the FOV of the VGA 17um Microbolometer when using the same lens. QED the VGA 17um microbolomter is present in both cases and down sampled/scaled in the lower resolution model.

The Ex5 model specs are available to compare here:

https://www.tester.co.uk/flir-e53-thermal-imaging-camera-24-degree-lens-240x180-20-deg-c-to-650-deg-c?gclid=CjwKCAjwrKr8BRB_EiwA7eFapi9fmQsz5luxU8HkocQoq9Fdzbe8NmKvYTHwe38yF00X8wbWvNB3OxoClIUQAvD_BwE

I know this does not answer your question but it may enlighten you as to why a VGA microbolometer is strongly suspected. Mass production can sometimes justify ‘overkill’ or use of QA ‘rejected’ microbolometers for lower resolution applications. I am not saying that such is definitely the case here though. One thing I do know is that FLIR will likely not want to comment in case it leads to an effort to unlock the full resolution of the Ex5 series cameras to full VGA resolution.

So why has no one already tried to hack the Ex5 series cameras ?

The answer may be that they are very expensive and often in the hands of professionals who have no interest in hacking the camera. A hobbyist would need to have deep pockets plus a large risk appetite to start digging around in the firmware of an expensive Ex5 camera. The Ex series E4 was relatively inexpensive and so people were willing to risk bricking it in order to discover its secrets. FLIR have learnt from past mistakes and intend to make life harder for those wishing to unlock a cameras full potential. They used Win CE6 in the Ex and Exx series cameras. That is an operating system that is very vulnerable to hacking. I was told that the Ex5 is running Linux but have not had confirmation of that. If it is running Linux, it can be made far more secure than Win CE and that could make hacking of that camera far more challenging. As I say, this has not been proven by me or confirmed for me by anyone at FLIR. From a personal perspective, a move to Linux would be a very sensible idea for FLIR but it would mean a significant change for its firmware development team who are used to Win CE.

Time will reveal the secrets of the Ex5 series but for the moment we have only comments on this forum and their sources may be reliable, or not, as the case may be.

So Calel, are you considering jumping from an E4 to an Ex5 series camera now ? Considering your concerns regarding the Exx series, that is quite a leap and one which will offer you no upgrade options at this time. Or is this just a casual question out of curiosity ?

Be aware that some of us on the forum have had a long term relationship with FLIR and their products so understand some of FLIR’s ‘habits’. We also receive snippets of information through friends at FLIR. We are not able to make all information public as that would be a breach of a confidence in some cases.

Fraser

« Last Edit: October 17, 2020, 03:13:47 pm by Fraser »
Cogito, ergo sum
 

Offline calel

  • Regular Contributor
  • *
  • Posts: 87
  • Country: ch
So Calel, are you considering jumping from an E4 to an Ex5 series camera now ? Considering your concerns regarding the Exx series, that is quite a leap and one which will offer you no upgrade options at this time. Or is this just a casual question out of curiosity ?
curiosity

though I am considering buying an old E40 if I can find one (to upgrade it to E60, if that's easy as some have...suggested). I saw the image quality differences between the E8 & the E60 and I'm...puzzled to say the least as to how 2 cams of equal resolution can have such a difference in image quality - is it all down to the fact that the latter has a manual focus?


Quote
Be aware that some of us on the forum have had a long term relationship with FLIR and their products so understand some of FLIR’s ‘habits’. We also receive snippets of information through friends at FLIR. We are not able to make all information public as that would be a breach of a confidence in some cases.

Fraser
hey as long as you dont rat out their names, you wont get them in trouble with their company so no biggie right?
 

Offline calel

  • Regular Contributor
  • *
  • Posts: 87
  • Country: ch
btw if FLIR use Linux on their better cams wouldn't that make it easier to hack? after all the source code of Linux is is available to the public which means people would know the innards of Linux way better than any Windows (whose source code is private), it'd make sense that if they know Linux better then they'd better know how to exploit it  ???


also how come Fluke IR cams are unhackable? did they take a different approach? (physical differences between different models of a same series?)
 

Offline mzzj

  • Frequent Contributor
  • **
  • Posts: 985
  • Country: fi
Fluke really do make awesome thermal imagers and I did always prefer them over FLIR.
Pretty much opposite experience here with Fluke thermal imagers: Weird issues with non-uniformity correction, many models are totally bonkers for first 15 minutes and firmware is really basic in the <5000USD "entry" level cameras.

20 years old NEC thermal imagers seem to outperform anything  under 10k usd today but NEC has never been budget-friendly option..
« Last Edit: October 17, 2020, 05:36:34 pm by mzzj »
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9909
  • Country: gb
You do not release information into the public domain that has been shared with you in confidence ! Doing so can set the hounds running within a company and information may be traced back to the source. In most cases the information shared with me is on the basis of trust that I will not put it up on a public forum. Breaking such a trust would result in loss of the source forever. Morals are important !

Linux is a sophisticated operating system used throughout Industry. It can be made very secure if the the author of a software desires such. Open Source does not mean non-secure .... it likely means better understood and more holes plugged in its design and security  :) Security needs to be built into an application rather than adding it as an afterthought. With physical access to the OS and it’s file structure it is possible to compromise an applications security, but it can be a very challenging task. With a thermal camera we are not talking about the type of security defenders found in Crypto machines but the security can still be tough to crack if the design team focus on the security aspects adequately.

Fluke cameras have not been cracked for the reason given above..... it was too much effort for a less popular, and often more expensive, camera. The FLIR E4 was a true gift in terms of the ease with which it could be upgraded in the early firmware versions, just like the Exx series. FLUKE use a better counter upgrade security model than FLIR employed and I have not found anyone wanting to spend the time cracking it whilst the FLIR offerings were still available and upgradable. FLUKE do the sameness FLIR in terms of common hardware platforms that are hobbled to provide the required performance and no more. I believe FLUKE to be more experienced in stopping users hacking their equipment to a higher specification.

Note that FLUKE have no true heritage when it comes to making thermal imaging cameras. They bought three companies that made ‘OK’ but not wonderful thermal cameras and used them to offer thermal imaging to those who will only buy a product with the FLUKE name.

Two of the companies were Raytek and IRISYS, the third names escapes me. I own a few FLUKE thermal cameras and they are far down my list of favourites, but that is just my opinion. I bought examples to investigate the firmware and method of hobbling applied to the resolution and functionality  ;)

The VT02 and VT04 come from IRISYS and the Ti30 came from Raytek. I think most of the current lineup of FLUKE cameras come from the third company ....... Infrared Solutions ? .... I will check

Fraser

« Last Edit: October 17, 2020, 06:10:56 pm by Fraser »
Cogito, ergo sum
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9909
  • Country: gb
Yep, my memory was correct, FLUKE bought Infrared Solutions in 2005. I never liked the offerings from Infrared Solutions...... they felt ‘cheap’ and looked ‘weird’.

Fraser
Cogito, ergo sum
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9909
  • Country: gb
Mzzj,

I agree regarding the NEC Avio cameras. I own some real beauties that produce superb radiometric imaging  :-+

No surprises that Agilent went to NEC AVIO when they wanted a thermal imaging partner to make annAgilent camera.

NEC AVIO are better known in Asia and are a very much respected and capable producer of high performance thermal cameras. The analysis software was not the best but they bought in “Thermography Studio” to address that weakness  :-+ NEC Avio cameras are certainly not ‘inexpensive’ but then real quality never is.

Fraser
« Last Edit: October 17, 2020, 09:59:02 pm by Fraser »
Cogito, ergo sum
 

Offline JimM

  • Contributor
  • Posts: 26
  • Country: us
I would like to see Seek step up and provide an improved Reveal (not necessarily the Reveal form factor), but an improved 320x240 resolution TIC with better accuracy than the Reveal in approximately US$1k price range.
 

Offline Bill W

  • Frequent Contributor
  • **
  • Posts: 541
  • Country: gb
    • Fire TICS
Yep, my memory was correct, FLUKE bought Infrared Solutions in 2005. I never liked the offerings from Infrared Solutions...... they felt ‘cheap’ and looked ‘weird’.

Fraser

Certaily wierd, but the ones I played with at a show in 2003 were not 'cheap' feeling, they were cast aluminium chassis.  Even then they were in Fluke yellow though !  Mind you that was the 'colour for builders stuff' then and now

Web page archive from the pre-Fluke era
https://web.archive.org/web/20031212181508/http://www.infraredsolutions.com/html/irinsight.html

You get images and text but not pdf.

Bill

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9909
  • Country: gb
Bill,

It was the Flexcam that I found cheap and weird feeling but the IR-Insight also had its issues. The yellow casing suffered from fatigue cracking over time :( I have worked on a few and all had the cracked shell. I own the IR Snapshot camera that is a sort of slow scan thermal camera with the sensor array moving horizontally on a sled. It looks weird, is weird and a bit of a beast  ;D It just seems to me that Infrared Solutions built ugly cameras that did not appeal to my eyes. It is a personal preference thing. The Flexcam looked like a direct ‘rip off’ of the old Sharp “Viewcam” Camcorder case design. The IR-Insight is a bit of an ugly spud as well. Basically a box mounted on a handle and it is very ‘brick on a stick’ in appearance. There just did not seem to be much effort in making attractive camera designs.

IR-insight brochure....

http://www.chevrierinstruments.com/Anglais/Nouvelles-002a/3-fluke-ir_insight-info-add.pdf

Fraser
« Last Edit: October 18, 2020, 12:00:06 am by Fraser »
Cogito, ergo sum
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf