Where next for American made miniature imaging core development ?

[Fraser]

I thought this topic worthy of some thought and maybe discussion on this forum.

If we wind the clock back to 2012 there was much 'chatter' regarding a new company that had been created to produce affordable thermal imaging cameras that would still offer decent imaging performance. That company was Tyrian Systems that we now know as Seek Thermal. That new company promised much as it was headed by very experienced and capable persons who had worked for Amber, Indigo and FLIR. The results of this companies efforts were a range of thermal imaging solutions from discrete cores through to complete cameras in either dongle or AIO format.

Around the same time, FLIR was developing its own new miniature thermal imaging core technology using the core miniaturisation expertise that they had acquired when they purchased Indigo Systems in 2004. The result of this R&D was the LEPTON miniature core that was integrated into a mobile phone 'back pack' that formed the FLIR One camera system

So that is the background to the USA's development of miniature thermal imaging cores that are public knowledge. The two companies continued to develop their miniature thermal imaging cores and i detail the products that resulted below:

FLIR

Lepton 2
Release Year: 2014
Resolution: 80x60

Lepton 3
Release Year: 2015
Resolution: 160x120

Boson
Release Year: 2016
Resolution: 320x256


Seek Thermal (was Tyrian Systems)

Seek Mk1 (Classic)
Release date: 2014
Resolution: 206 x 156

Seek PRO Mk1
Release Date: 2016
Resolution: 320 x 240

The above list does not detail the later revisions of each core as they have not changed significantly in design.
The FLIR BOSON is available in a 640 x 512 pixel version but is omitted from this list.


What is apparent from the details provided above, is that the USA based commercial miniature thermal imaging core development is in the Doldrums ! The Lepton and Seek Thermal cores were first released in 2014, with resolution enhanced versions being released soon after in 2015/2016. Since then we have seen the release of the Flir BOSON and some minor development work on the SEEK thermal cores to reduce the internal noise levels to acceptable standards. There has been no significant improvement or development in the miniature thermal imaging cores coming out of the USA in recent years. This leads one to wonder whether these two core manufacturers have decided to concentrate on recovering their development investment before releasing new products or are they truly sat in the Doldrums and not developing more advanced miniature imaging cores. None of the cores detailed in my list above could be called 'cutting edge' any more or even "high performance". The designs are getting long in the tooth and China based microbolometer makers have the miniature thermal imaging core market firmly in their cross-hairs.

What about the FLIR BOSON, some may ask.

I included the BOSON in my list because it was the Darling of FLIR that promised much going forwards in the market. Yes it was more expensive than the Lepton, but it was smaller than the Tau series and effectively replaced the MUON and QUARK series in FLIR's core lineup. The BOSON was developed and released around the same time as the LEPTON 3 core. It promised to be a solid platform on which to build a new and highly capable range of thermal imaging cores to slay the competitors in the OEM core market segment. There was a problem however..... the early tests on the Boson showed it to deliver underwhelming imaging performance The BOSON clearly had problems but any corrective action was to prove a challenge due to FLIR's use of a very advanced MOVIDIUS image processor with which its engineers were not yet fully comfortable when it came to making best use of its powerful image processing engine. Without going into details here, the Movidius processor and its configuration still presents a significant challenge to FLIR. The BOSON contains a processor that has been shown to have amazing capabilities, yet the BOSON still fails to impress. I am sad to say that I believe the BOSON, with its MOVIDIUS processor, to be a development that will wither and die, rather than drive future generations of imaging core coming out of the FLIR R&D department. It was a great idea but, IMHO, it failed at the implementation stage.

So here we are in 2021, looking at the miniature thermal imaging core options from The USA, complete with their frame rate restrictions to meet the current US regulations for this technology. They are looking a little old in terms of their imaging performance.

China based companies developing their own microbolometers have not been sitting on their hands ! The Worldwide Pandemic saw a huge push from the Chinese Government to develop thermal imaging technology for use domestically and overseas. We have seen a significant increase in thermal imaging camera solutions appear on the market in the past two years. .The Chinese microbolometer manufacturers have produced new miniature imaging cores that deliver on imaging performance and price. The Camera OEMs have been sourcing domestically produced thermal imaging cores without the constraints of US Exports 'red tape' or the complications of the language barrier when it came to product development using the cores in 3rd party hosts. These thermal imaging solutions are also shipping with frame rates of 25fps which make the US supplied cores appear inferior when the camera system is in motion. the resolution of these new China made cores is interesting. The traditional resolution steps of thermal imaging cores were 80x60, 160x120, 320x240 and 640x480 pixels. To this we saw the 384x288 pixel cores added and SEEK Thermal used the 206x156 pixel format in their standard cores. The Chinese sourced thermal imaging cores are commonly available in the standard resolutions but to these we have recently seen 120x90 and 256x192 added. The 256x192 pixel format is of interest as it sits between 160x120 and 320x240 so is a sort of compromise solution....not low enough resolution to cause pixelated images, yet not so high a resolution that it needed a large microbolometer die with the associated higher production costs.
Someone clearly thought about the efficiencies offered by the 256x192 pixel microbolometer and the evidence collected from cameras that use this resolution of core shows it to be a very pleasing compromise. Enough detail is present in the images for many applications, including PCB inspection.

So what future for USA designed and produced miniature thermal imaging cores now that China has most definitely entered the market with decent products. They offer affordable prices to OEM's and the added benefit of easy international shipping and 25fps frame rates ? FLIR and SEEK Thermal have yet to show any signs of responding to this threat to their market segment, beyond FLIR reducing the cost of its elderly Lepton cores. They need to react soon or the market will move on and adopt the new Chinese cores for their new product releases. Many Chinese thermal camera producers have already moved away from FLIR and SEEK Thermal cores and having read this post, can you blame them ?

As end users of thermal imaging products, we are benefitting from the release of these new thermal imaging cores coming out of Chinese factories and we may yet see a strong response from the Wests Microbolometer and thermal camera manufacturers. The alternative for FLIR would be to effectively abandon the consumer product market and concentrate on its Industrial and Military product lines. Sadly SEEK Thermal have no such alternative income stream and I can see tough times ahead for that company. Seek Thermal are competing with China based manufacturers who produce a product that offers better imaging performance at lower cost and with a higher frame rate.

Food for thought and all my own personal opinion so make up your own mind 

Fraser

[SilverSolder]

Not a problem, we can just ban the Chinese products to protect our own manufacturers, and just live with slightly less performance - who cares if it isn't the best, it's made by our boys and girls!!! 

I shouldn't joke about it...  this approach seems to be the clear political trend in recent years....

[Fraser]

In the past, competition was good as it was a driver for lower prices and increased development. I have not see this happening with FLIR or SEEK though. It is as if they have their heads buried in the sand. Could they have been caught off guard by the new core releases from Infiray and Guide Sensmart ? If so, they need to improve their industrial intelligence gathering techniques !

I tested and reviewed the DYT CA10 PCB analysis system and was very pleasantly surprised with the imaging performance being provided by the Chinese made miniature thermal imaging core. The core is about the same size and format as the SEEK Mosaic with which it competes. Not by chance me thinks ! DYT used to use the SEEK Thermal PRO QVGA core imported direct from the USA with the export controllers blessing. They have moved to a domestic core manufacturer and gained easier supply and better technical support.
DYT put the effort into designing a high quality mounting system for the camera head and needed a decent imaging core to attach to it. 2 years ago they would have used the SEEK Thermal PRO, I have no doubt about that as that was all they could source. Now they had the option to use a core from a manufacturer who was falling over themselves to supply what was needed at a good price. The result is a very nice PCB inspection camera that I prefer to my FLIR ETS320 in terms of its excellent ergonomics and manual focus.

I am currently testing and analysing a Hikvision DS2TP31B Fever Detection camera that provides a 160 x 120 pixel image. I did not find a FLIR lepton inside it though. I found a Hikvision microbolomter that uses the same technology as the FLIR E4 camera with a directly bonded MEMS microbolometer on the sensor PCB. It is only 160x120 pixels though….the microbolometer core offerings from Infiray are, IMHO, superior.

Fraser

[Bill W]

It is quite likely that FLIR are in corporate paralysis after the takeover from Teledyne.  That in itself may kill off the consumer side.

Meanwhile, although FLIR have got bulk 30Hz export approvals for professional markets anyway, Teledyne customers do not need to worry about US export controls - there was already a pure Teledyne camera and sensor line in Canada running under conventional Wassenaar rules.

Bill

[jhhong98]

Excellent review, a veritable State of the Thermal Imaging Market analysis!   The entire industry raced toward smaller pixels not as a way to improve performance (noise worsens and the gain drops in proportion to the pixel area, offset only by using faster lenses) but presumably to cut costs.  But the prices have not decreased significantly.  Watch this CES presentation given together by Veoneer and FLIR.  The self posed question is why thermal imaging is not used widely in automotive and they answer that it is the price, BUT they are improving.     All of the questions from the audience are "why is it so expensive?"  The entrance of chinese makers have not really impacted overall prices.   I would say most VGA cameras go for $2000+.   Your assessment is spot on.   

[bap2703]

I think everybody is just reaching the same limits of the technology. So there's only small margins for improvement cost wise and performance wise.

Right now there's probably more value to be found at the system level and I guess that Teledyne will focus on more refined higher end systems. In these the cost of the hardware can be diluted because they are offering much more than just a sensor.

The industry has a safeguard: domestic microbolometer production will never cease (unless there's a better alternative of course) due to it's importance in heat seekers and other military thermal vision equipement.

So for what we're looking for here on this forum: sadly I don't think we'll have affordable fullHD thermal devices any time soon.