EEVblog Electronics Community Forum

Products => Thermal Imaging => Topic started by: Ultrapurple on February 28, 2020, 07:43:06 pm

Title: SEEK Reveal Pro FF review by Ultrapurple
Post by: Ultrapurple on February 28, 2020, 07:43:06 pm
SEEK is a relatively new name in thermal imaging, though the people behind it have a long and distinguished track record, documented elsewhere (https://www.eevblog.com/forum/thermal-imaging/seek-shot-series/msg1949125/?topicseen#msg1949125) on this forum.

The first Seek thermal cameras were relatively low resolution dongle types, now known as the Compact series. These offered 206x156 resolution and in my opinion, at least in the earliest types, had quite poor performance. I believe others have mentioned also that there was a thermal gradient issue across the sensor, though I rehomed my Seek Compact XR before using it enough to spot that.

The next product range was the all-in-one Reveal series, which combined a 206x156 resolution sensor with a battery and LCD display in a hand-held package. These were nice to handle but again I believe the resolution and sensitivity were limiting issues. Slow (<9Hz) export versions gave jerky results, as one would expect; the >15Hz US versions were better in that regard.

[attachimg=1 width=320]


Further development of the Seek sensors resulted in a remarkably good 320x240 pixel unit. It is offered in various form factors (dongle, OEM module or all-in-one camera as seen above). I’m looking at the all-in-one Seek Reveal Pro here, in the Fast Frame (>15Hz) version. The Fast Frame variant is easy to spot because of the ᖷᖴ 'FF' logo on the top left of the device. The <9Hz version has no such marking but is otherwise outwardly very similar.

The Seek Reveal Pro is an all-in-one high resolution thermal imaging device that fits nicely in the hand. It has something of an air of Star Trek about it. Its sculpted form is thicker at the camera end and the camera is angled slightly downward so it looks straight forward when you hold the screen at a comfortable viewing angle. Ergonomically, it is a superb piece of work. I invited some colleagues to play with it. They have never shown any interest in my various other thermal cameras but this one garnered a great deal of attention. That says a lot about the quality of the design.

Here it is with an old Garmin etrex Legend hand held GPS and a Therm-App thermal camera for comparison.

[attachimg=5 width=320]


In order to make the device fit the hand so nicely, Seek decided to orient the camera and screen in portrait mode. I generally abhor ‘vertical video’ but in the case of this device it makes sense to use portrait mode and the format works very well. The display claims to be 320x240 pixels – a 1:1 match with the sensor – but the bezel masks a varying proportion of the screen so I guess there are no more than 220 pixels viewable at the ‘bottom’ of the screen. In practice this has no effect on usability. The following image shows an off-screen photo (left) and a saved file (right). You can see that the bottom of the screen crops the image somewhat, but not enough to be troubling. The moire pattern on the photo is an artefact of the photography and there is no such unevenness on the actual display. That said, the LCD has a slightly visible 'shimmer' to it as it refreshes from bottom to top. It is noticeable if you look for it but not intrusive.

[attachimg=3 width=320]

The Seek Reveal Pro comes in two varieties: regular (<9Hz) and Fast Frame (>15Hz) variants. I found a brand new FF version on eBay for under £500 (though the list price is nearer £700). The temptation was just too much!

It arrived after a couple of days, factory sealed with “Subject to US EAR” security seals. It’s worth knowing that I was not asked to sign anything to accept these restrictions (or even notified of them by the seller). The device has similar wording emblazoned on it and also shows it on the boot screen.

The Seek Reveal Pro is incredibly easy to use. You simply press the centre button to power it up, and thermal images appear within 10 seconds. A faint tick can be heard each time the FFC flag moves into position, but unless you're in a silent place you'll have difficulty hearing it. As is normal, FFC happens every few seconds after the camera is switched on but get further and further apart as the camera temperature stabilises. After several minutes there are long intervals between FFC events.

The user interface is just three buttons mounted above the display (or four if you include the torch button on the side). The central button acts as the power switch and also calls up the menu. Palettes are easily selected from the top level by pressing the left button. As with all menus, the left and right buttons scroll down and up, then the centre button selects the highlighted value. It's simple and intuitive (with the possible exception that the different palettes are selected using the ‘filter’ option). All the defaults are sensible.

Perhaps the most important thing is that the software is very snappy, with no lagging. Saving an image by pressing the right hand button is almost instantaneous. This is in stark contrast to other Seek cameras, which can take a couple of seconds to save an image. About 4GB of non-expandable internal memory is provided, which is accessed as a memory stick via the micro USB charging port. PNG (non-radiometric) images are about 225kB and radiometric TIFF images are about 750kB, so that 4GB will store a very large number of images (well over 4,000). I was pleased to discover that my phone is able to access the images by using a simple USB connection.

I am informed that it's vitally important you never try to format the image memory. Deleting images is fine, but the physical memory device inside the camera is partitioned so part is the user-accessible image memory and the remainder is system memory used to store firmware etc. Formatting the memory runs the risk of wiping the firmware and bricking the device. (Thanks Fraser).

[attachimg=2 width=320]

The most amazing thing for me was to learn that the Seek Reveal Pro is radiometric. A central reticle can be switched on; the temperature of the spot is shown in centigrade, fahrenheit or kelvin (as per your setup options). Without access to a blackbody source I couldn’t confirm the accuracy but all the readings I’ve seen so far looked pretty realistic. Bravo, Seek!

[attachimg=15 width=320]

[attachimg=4 width=320] 

Battery life is superb. SEEK quote four hours but I managed nearer five. It depends, of course, on how bright you run the screen. I found that setting it to 80% brightness was usually plenty, even outdoors. The 300 lumen flashlight can be set to two different levels. The defaults are something like 90% and 6%. At 90% it is blindingly bright!

Now, the most important thing: image quality. Quite simply, brilliant. I couldn’t believe that I had paid under £500 for such a good picture. There’s virtually no lag; the screen is bright; the FFC flag interrupts are fast and not particularly intrusive; and it’s an absolute joy to use.

Sensitivity is probably in the region of 70mK. That’s not as sensitive as the devices based on Ulis sensors and germanium optics (Therm-App, ThermalExpert etc), but it is certainly more than good enough to visualise small heat leaks in buildings. Once you’re into larger thermal contrast situations the differences between cores tend to become irrelevant anyway. There is very little (if any) fixed pattern noise. The image processing really is very good indeed. The camera copes incredibly well with extremely high contrast scenes that some other cameras have trouble with, for instance it can show cloud detail (effectively ~ -40C) and a hot building (a roof may heat above +40C) in the same image without any difficulty or black- or white-out. (Thanks to Fraser for drawing my attention to this).

I did a side by side comparison of the Seek Reveal Pro, a Therm-App (384x288) with 13mm f/1.0 lens and ThermApp Plus (640x480) with 19mm F/1.1 lens. The Therm-App cameras are definitely a bit more contrasty but (especially in the case of the $3500 Therm-App Pro) we’re not comparing like with like. The Seek has a relatively small chalcogenide glass lens whereas the Therm-Apps were using germanium. I haven’t been able to find out much about the Seek lens but I suspect it’s around  f/1.2. It’s fixed focus, unlike the others I mention, but set up nicely so that everything from a foot to infinity are in good-enough focus.

Left to right: Seek Reveal Pro FF (320x240), Therm-App (384x288), Therm-App Pro (640x480), all sized to 320x240 but click for original size. Note how the handles on the filing cabinets can't be seen in the Seek image but are visible in the Therm-App images.

[attachimg=7 width=320]   [attachimg=8 width=320]   [attachimg=6 width=320]

The sensor on the Seek Reveal Pro is made with 12µm pixels, which are some of the smallest on the market. (Therm-App and most other contemporary uncooled microbolometer thermal imagers uses 17µm pixels, whether they are VOX or A-Si). The benefit of smaller pixels include a physically smaller (=cheaper) sensor and a smaller lens because it doesn’t have to create such a large image circle as required by bigger sensors. The downside is that the pixels are only half the area of conventional 17µm sensors, which means there’s less sensitive area. All other things being equal, that means the pixels will be less sensitive than their larger brethren. There’s also the consideration that the pixels are nearing the practical limit for what will actually work in the 7-14µm wavelength range: they are less than one wavelength of light across.

Image resolution is 320x240 pixels, but that may not be the whole story. Particularly when pixels are small in relation to the wavelength, very great care must be taken with the design of the optical system (lens) in order to maximise the spatial resolution. The details are highly technical and Seek don’t release much information about their lens but, all things being equal, I would expect the optical resolution to be less than one would get with a comparable lens on a larger sensor. That said, it’s pretty darned good: Seek clearly have excellent designers who know exactly what they’re doing.

So far so good: everything has been overwhelmingly positive. Are there any downsides? Well, although the pictures flow nice and freely there is no video recording mode nor any way to extract a moving image signal. That could be a deal breaker for some. Otherwise I can’t think of anything else that might be considered a problem. Mike spotted space for a Wi-Fi or Bluetooth module on his teardown video so there may be a new version to come that includes such a facility.

My overall conclusion is that the Seek Reveal Pro Fast Frame is the best budget thermal imaging camera I have ever seen. It weighs next to nothing, is robust, fits in your pocket and gives superb images. I have no hesitation in recommending it.

(Note that these remarks apply only to the Seek Reveal Pro Fast Frame; other variants, such as the Reveal / Reveal XR use a different sensor).

Sample images - click on the thumbnails below. All are as-shot (except the landscape format images, which were turned in an image editor but otherwise not changed).
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Ultrapurple on February 28, 2020, 08:19:57 pm
Some features I forgot to mention. There are three basic display modes: normal, 'full frame' and Span.

Images are normally stored as non-radiometric PNG. A menu option lets you save all temperature data, in which case the image is stored as a radiometric TIFF.

More sample images:
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on February 28, 2020, 09:26:10 pm
Thanks for the review and pictures Ultrapurple.

I am tempted but just can not justify such a purchase at the moment.

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Ultrapurple on February 28, 2020, 10:42:21 pm
I couldn't justify it either.

It's bad enough being impecunious; one can hardly be expected to compound the misery by denying oneself the occasional treat...
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on February 28, 2020, 10:52:07 pm
Thanks @Ultrapurple

Gee, this has got to be the most "glowing" review of a Seek product I have seen on this forum!

But before anyone gets too excited, can you show us a photo of a low contrast scene with the entire gray scale spread over a ~3 degree C span?  ...Or even a 3 degree Fahrenheit span.   :D

Given the low cost business model I expect the lens is the same as on the Compacts.  Check the aperture; if it is 5mm & the FOV is the same as the Compact Pro then it is likely F/1.4

Fortunately you corrected your file size typo before I got logged in.  ;)

Anything from Seek on how to get the data out of those TIFFs (which I already did)?
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Ultrapurple on February 28, 2020, 11:48:12 pm
@IwuzBornanerd thanks - yes, I am aware it's a very positive review. I considered the camera from a position of experience with many other thermal cameras and it is impressive, offering an unmatched price/performance ratio. There is no doubt that there are better small form-factor cameras out there - for instance some based on the FLIR or DRS cores - but the cost of those is significantly higher (×10 or more in many cases) and none come in such an ergonomically pleasing package.

Perhaps we will see even better cameras at lower cost from the Far East in due course. Meanwhile I think this is the all-in-one to beat.

As I mentioned here and elsewhere, I was not at all impressed with the earlier Seek devices. I wonder if they will leverage their present sensor technology to make 640x480 or even higher resolution sensors in the future? It will be exciting to see.

As soon as I get the chance to set up a suitable scene I will make some images in a low-span environment. Most people are aware that low-contrast scenes are the 'torture test' for thermal cameras and it will be interesting to see how it compares with other devices. I am particularly keen to compare it to my FLIR SC-660 science-grade camera! The Therm-App cameras also provide a useful benchmark.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Ultrapurple on February 29, 2020, 09:23:00 pm
Now for some relatively unscientific test results. I compared the Seek Reveal Pro against
I chose four scenes for the tests and made the images as close to each other (in time) as practical. All of the cameras were allowed to warm up for at least five minutes before any images were made and (in the case of the Therm-App cameras) a manual FFC was performed immediately before the images were made. The FLIR and SEEK devices have automatic FFC.

The scenes are:

The FLIR SC-660 is a high class science grade camera. Although it is not in calibration, a recent check against a good blackbody source (thanks, Fraser) indicates that its temperature indications are within about 1°C of correct, so this is what I'm using as the 'definitive' reference standard.

The following results are scaled to a total of 640 pixels wide; click on the images to see them full size.

Results from FLIR SC-660:

[attachimg=4 width=640]

Results from Therm-App Pro:

[attachimg=1 width=640]


Results from Therm-App:

[attachimg=2 width=640]


Results from SEEK Reveal Pro:

[attachimg=3 width=640]


The results are fairly clear. The minimum span torture test of the brick wall is, realistically, beyond the capabilities of the SEEK Reveal Pro. That shouldn't be seen as too surprising, as the thermal contrast is probably well under 2°C. The SEEK fared much better when looking at the PVC door (the patterning is a reflection of the street scene) and by the time there was some real contrast (on the path), the SEEK's results were very acceptable. Not surprisingly, the wide temperature span of the cars image gave pretty good results on all of the cameras.

I have done numerous other tests on the SEEK that indicate it is sensitive enough to see quite low-level heat leaks in a domestic environment, though it's clearly not as good as the other cameras.

Conclusion: as per my earlier notes, the Seek Reveal Pro isn't as sensitive as the higher-priced alternatives, but in most real-world conditions it acquits itself remarkably well.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Ultrapurple on February 29, 2020, 09:36:32 pm
Here are some original image files from the different cameras - the FLIR SC-660 (IR... file names) and Therm-App (T A file names).
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Ultrapurple on February 29, 2020, 09:37:14 pm
... and the SEEK and Therm-App Pro (TA Pro) files:
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Ultrapurple on February 29, 2020, 09:48:57 pm
More files from the SEEK Reveal Pro:

Small dormer window showing battens where the internal panels are fixed

[attachimg=11] 

Air conditioner

[attachimg=1]

Flex of a 3kW electric kettle and mains plug

[attachimg=2]

Kettle, just off the boil

[attachimg=3]

Vent from a portable refrigerator, seen in different palettes

[attachimg=4]  [attachimg=5]  [attachimg=6]  [attachimg=7]  [attachimg=8]  [attachimg=9]  [attachimg=10] 





Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on February 29, 2020, 10:16:02 pm
Thanks for the comparison and pictures Ultrapurple.

Considering what I remember of my original SEEK phone dongle, this Seek Pro core is vastly superior in terms of image noise. Cameras using the 12um pixels (actually only 10um receptors)  have to work very hard to produce a decent low noise image of low Delta T scenes. I can forgive the Seek Reveal Pro for not excelling in that part of the test.

I would love to see how the SEEK Reveal Shield Pro performs as that unit appears to be a ‘tuned’ version for law enforcement use and it has a very large lens compared to the other models. Sadly there is little chance that I will get my hands on that model though.

https://www.opticsplanet.com/seek-thermal-reveal-shield-pro-thermal-image-camera-rq-lah.html (https://www.opticsplanet.com/seek-thermal-reveal-shield-pro-thermal-image-camera-rq-lah.html)

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Ultrapurple on March 01, 2020, 12:21:30 am
@ Fraser - If you take a look at the Seek site (www.thermal.com (http://www.thermal.com)) you will find that they offer two types of tiny OEM cores with different lens options. The 'Mosaic' cores (https://www.thermal.com/uploads/1/0/1/3/101388544/mosaic_core_specification_sheet.pdf) are available in 320 x 240 resolution with a variety of lenses including a 9.1mm f/1.0 that looks suspiciously like the one on the Shield Pro. That would give very long range detection (Seek say 750m) and good sensitivity. Other lenses in the series range from f/1.05 to f/1.26. However, none of the listed options quite match the Reveal Pro so I don't know what focal length or aperture it has. (Knowing the pixel size, FoV and measuring the lens entrance pupil diameter should give some clues but I haven't done the sums yet).
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Ultrapurple on March 01, 2020, 01:04:13 am
I wonder what would happen if one grafted a different lens onto the sensor? Something like one from an Argus 1 or 2 would give excellent long reach and, as the aperture is quite wide, the images would probably be nice and contrasty. Of course, it would mean a severe risk of wrecking the camera and even if it worked the calibration would be way off, but still...

Actually, I guess there would be more issues than just calibration. Even assuming something could be contrived that didn't foul the FFC flag, I suspect Seek is doing something clever to compensate for fall-off at the edges of the image circle: cost engineering dictates that the image circle should be as small as you can get away.with so you economise on lens materials but if it's too small you get objectionable vignetting in the corners - unless you carefully tune the gains in the corners to compensate. You would get more noise in the corners of course,  but normally one is mainly interested in the centre of the image. Just a thought...
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 01, 2020, 01:20:44 am
Ultrapurple,

I did some reading of the Reveal Shield Pro user manual to see if they detailed anything interesting compared to the standard Reveal Pro. The Shield has a 24 degree HFOV which is a very common figure for professional cameras used in Industry. The only significant differences that I could find were the clearly larger lens and three search specific LUT’s. The unit defaults to ‘night’ mode with a low intensity red monochrome image. This is intended to preserve night vision and reduce the users face illumination. SEEK suggest that this is the default mode to protect officers in high risk situations. There are two other modes, one monochrome and one colour. They are both high contrast modes intended to highlight suspects in a scene.

I was wondering whether the Wi-Fi module might be present in the Reveal Shield model for video streaming or image download, but no, there is no mention of it in the user manual.

To be honest, I would not personally use a thermal camera of the Reveals format for law enforcement searches. In the covert mode of observation you want as little of your body visible as possible and standing up with the Reveal at waist or chest height viewing a scene exposes a rather large target if the quarry is armed. A thermal scope such as the FLIR Scout and similar format units plus a military iris type eye cup is my preferred tool for the job. The Reveal Shield might prove very useful in non covert roles though. Searching for missing persons who present no threat comes to mind.

I would still like to see how much difference that larger lens makes to the camera cores noise levels though.

Oh I almost forgot, the Shield is available in <9fps and 30fps versions. Yes 30fps ! However the UK has had Seek Reveal XR30 units for a while now so I knew there were 30fps versions around.

As a side note, these very compact thermal imaging cores might provide improved imaging performance if a better lens were to be fitted. Such an exercise would involve modifications to the camera but might give a worthwhile improvement in performance. The poor little FLIR Lepton is a classic case and would almost certainly benefit from a nice good quality Germanium conventional lens in place of its silicon diffractive optics. I have some very neat little high quality Germanium primary lenses but would I want to carve up my nice new SEEK Reveal to use one on it ? Hmmm now there is a question.

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 01, 2020, 01:31:40 am
Ultrapurple,

I just saw your post.... our minds think alike  ;D

I was thinking more along the lines of a Germanium lens from an early model Argus 4 or MiTic camera  ;)

There may be some noticeable difference in the lens characteristics effecting the NUC and measurement calibration but you might get lucky or the improvement might make such still acceptable. The only way to find out is to get myself a sample SEEK camera and test the theory. Not at the moment though as I am resisting further thermal camera purchases. With regard to using a lens assembly from a somewhat elderly thermal camera such as an Argus 1 or 2, there is the issue of resolution to consider. The SEEK pixels are tiny compared to the monster sizes used in the Raytheon BST FPA. IIRC, the BST pixels were 50um but I may be wrong about that. Such large pixels will likely be far less demanding of the lens system illuminating the FPA. Still good quality Germanium lenses, but likely not specified for the tiny 12um pixels.

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Bill W on March 01, 2020, 11:11:36 pm
Quote from: Fraser on March 01, 2020, 01:31:40 am
I was thinking more along the lines of a Germanium lens from an early model Argus 4 or MiTic camera  ;)
With regard to using a lens assembly from a somewhat elderly thermal camera such as an Argus 1 or 2, there is the issue of resolution to consider. The SEEK pixels are tiny compared to the monster sizes used in the Raytheon BST FPA. IIRC, the BST pixels were 50um but I may be wrong about that. Such large pixels will likely be far less demanding of the lens system illuminating the FPA. Still good quality Germanium lenses, but likely not specified for the tiny 12um pixels.

Fraser

Indeed, also the Argus 1 lens is f/0.7, as a tube camera was essentially working as around 100µm pixels, with 200 lines over 18mm.

Some brief lens specs:
These are the spec limits, so a typical lens will comfortably exceed these.

One of my 'things to do' is to make some Frankenstein cameras putting various lenses onto modern cores including a FLIR One.  I may even end up making adaptor plates.

Bill
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on March 02, 2020, 12:57:23 am
Ultrapurple,

Don't want to hijack your thread, but was curious -- are the Flir images straight from the camera, or did you go through Flir Tools?
I am attempting to reconcile the temperatures shown at the center spot with my calculations for temperature. I get slightly different results.

Thanks.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Ultrapurple on March 02, 2020, 01:13:01 am
@ BillW - I love the idea of a Frankencamera!  :D

@ JimM - no problem. All the FLIR images are straight off the camera. It looks as though I inadvertently used two different versions of the first image because the one in the composite image has crosshairs and temperature figures but the separate file doesn't. I didn't use FLIR Tools (I don't have a licence code for it).

The IR... jpgs are the FLIR radiometric images so feel free to use them to see if my camera's mileage differs from yours. There are some more SC 660 radiometric images here:
https://www.eevblog.com/forum/thermal-imaging/flir-sc-660-vs-therm-app-pro-thoroughly-unscientific-comparison/ (https://www.eevblog.com/forum/thermal-imaging/flir-sc-660-vs-therm-app-pro-thoroughly-unscientific-comparison/)
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on March 02, 2020, 10:07:10 am
 Excellent comparison Ultrapurple; thank you.  That is the way to compare cameras.  Great to see same-scene comparisons from different cameras.

The thing I find the most amazing is the lack of difference between the images of the pavement.  I wonder if the texture of the pavement masks some of the Seek noise.

I notice that the Seek chose a 2 degree span (thus showing more noise) for the brick but the FLIR chose 4 degrees.  Do you know what the ThermApp span was?

It looks like you have the Seek image smoothing off, (judging by the door image particularly).  Is that true?

I suspect images with the standard color palettes will show as much noise with around 2x the temperature span as the gray scale.

For comparison, I'll post a shot from my FFPro with 3.2C span & fully warmed up & fresh 2-point NUC (and no smoothing).  Looks pretty much the same as your door shot noise-wise, although I can't duplicate your scenes.  I make note of this as a way of saying that the Reveal Pro & Compact Pro appear to have the same image performance.  The Reveal may have more stability & repeatability as a result of it's packaging, though.  I wonder if your evaluation will help remove the stigma associated with Seek that resulted from the original units.

The TIFF posted off the Seek app by another user has the temperatures with 5 decimal places, apparently 1/32 degree increments.  Given that, it should be possible to take any image & zoom in on a particular temperature range & see the noise regardless of what span the camera software was using at the time.  If you'd care to post one or 2 of those TIFFs off your Reveal Pro (maybe narrow & not so narrow span) I'll see what I can do with them.  :)

I forgot to mention that I have on my "to do" list getting some form of F/1.0 lens on one of  my Compacts.  As best I can determine from measurements, the "tunnel" in the Seek Compact lid will support an F/1.0 aperture but the bayonet will not.  And about the only focal length I could get F/1.0 in with the cheap ZnSe lenses would be half inch--by using 2 1" focal length plano-convex lenses.  I THINK this can work on the Seek but it might not be possible to focus at long distances.  Going to F/1.0 would double the signal-to-noise & thereby hopefully get the Seeks down to 1LSB of noise.  I figure they started with the 4 LSBs noise & are now at 2 bits of noise.  With F/1.0, though I'd need to figure out how to get the temperatures right.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on March 02, 2020, 07:10:13 pm
Quote
The IR... jpgs are the FLIR radiometric images so feel free to use them to see if my camera's mileage differs from yours. There are some more SC 660 radiometric images here:
I don't actually own an IR camera...yet. I got interested in Thermal imaging a few months ago from a software development standpoint -- first from the ExifTool forum and then here. Comparing FLIR temperature data from radiometric images posted online, sometimes I get exactly the same temperatures, other times they differ a little. I am basing my temperature calculations on the equations that tomas123 posted in a spread sheet several years ago. I was attributing the differences to possibly the use of FLIR Tools, since in FLIR Tools you can tweak things like emissivity, etc. However, you don't use FLIR Tools, so I can eliminate that possibility. Will take a look at your other SC 660 images. Thanks again for your help.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Vipitis on March 04, 2020, 02:30:12 pm
While almost off topic:

Bill - I managed to put the 18mm f/0.7 Lens I got from you a while ago onto my phone's camera which is a Lepton 2 core, very close to the FLIR ONE spec. I sadly broke one element so I can't test it on my bigger(but older) cores with direct sensor access. The tiny 4.5mm prototype did not cover the Lepton fov at all and removing the integrated lens wasn't an option for me. I will gladly share my CAD model for the lens adapter I printed with my measurements for changing the distance between groups. I couldn't get any results for resolving power.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 21, 2020, 04:43:09 pm
I sold a few items on eBay to finance the purchase of a Seek Reveal Pro Fast Frame model  :)

In the past I have been less than impressed with the early Seek Thermal Classic dongle. I sold mine soon after buying it. Since then I have watched with interest as the Seek Thermal products image quality appeared to improve. Noise levels have been lowered and the QVGA Pro core produces decent imagery without the need for fancy post capture PC/MAC image enhancement. I recently asked this forum whether this was indeed an improvement in the cores, as I suspected. It became apparent that Seek have been working on the noise levels present in their systems and also the noise processing elements of their firmware/software. Was it time to dip my toe back into the Seek Thermal waters I wondered.

A pair of friends (fellow forum members) showed me their Seek Thermal Reveal cameras. The first one I saw was a standard Seek Reveal and the second unit was the Seek Reveal Pro Fast Frame belonging to Ultrapurple.  I was pleasantly surprised at the picture produced on the standard Reveal unit and began to question my somewhat negative view of the Seek Thermal products. Ultrapurple very kindly visited me with the Reveal Pro Fast Frame and I was further impressed by the improvements that I saw in imaging quality. Let’s not get carried away though, we are still dealing with a core that struggles to produce a low noise image due to the challenges created by its diminutive pixels that measure barely 10um in true pixel size. The question is, have Seek Thermal done enough to tame the noise to acceptable levels for most users ? My opinion is that with the Seek Reveal Pro they have succeeded in this quest  :-+

The Seek Reveal Pro Fast Frame is quite an expensive item, even though cheap compared to other complete camera systems of the same resolution. Ultrapurple and I had seen an eBay advert by “Yoltso” for the Seek Reveal Pro FF model at a lower price of £489 in the UK. We both thought the price suspicious, especially as the same seller had listings of the same camera at higher prices ! Was it a scam or reconditioned stock we wondered. We independently researched Yoltso and they are most definitely legitimate. Kondor is another name associated with them and they have huge supply contracts with Vodaphone. The price still seemed too good to be true though. Could it be a typo in the listing or incorrect model ? Was it truly a Fast frame model. I was negotiating with another private seller who was asking £500 for a used Reveal Pro FF and he did not believe the Yoltso price and thought it a non FF version. Well as we know, Ultrapurple bought a Yoltso supplied Reveal Pro FF and it was brand new, sealed in the box and the Fast Frame version. I have no idea how or why Yoltso are cheaper than any other UK supplier and effectively cheaper than the USA when UK VAT and Camera Duty is added.

The bad news was that I did not have the £489 spare so had to resist purchasing the Reveal Pro FF. I was considering buying the lesser Reveal standard model instead but held back as I knew I would always want the better specification image and menus etc. Not long after this I restarted my “Grand Clearout” of test equipment on eBay and that has gone well. With all the trouble in the World I decided to give myself a present from my earnings  ;D I took another look at the Yoltso listing for the Reveal Pro FF and suddenly saw an eBay 24 hour limited offer pop up...... Spend over £100 and receive 10% discount up to a maximum of £50. It was a sign ! The time was right, the planets were aligned in my favour and the stars shon on me at 01:15 in the morning ! I pushed the “buy it now” button, applied the 10% discount code, and the Reveal Pro FF was mine for £440  ;D

The camera arrived via DPD and came a day earlier than predicted. The innocent little padded bag got past my wife’s toy detector without comment :) I had to go shopping yesterday but upon return I unwrapped the precious cargo in our bedroom, away from prying eyes  ;D The Cameras box was new, fully sealed and its contents perfect.
You hold that finely crafted Reveal Pro casing in your hand and it just feels ‘right’  :) The charge level was at 70% so it is not old stock. I could find absolutely no reason for the unit being so discounted by Yoltso.

About the Reveal casing and design..... fans of Star Trek will love it ! The casing is, quite frankly, a thing of beauty. The dimensions and curved shape are so well considered for a comfortable grip. The PRO model has the ‘carbon fibre’ pattern effect on the plastics but it does not look cheap as can be the case with some fake carbon fibre effects. The unit is so solid and has an air of quality exuding from it. Tough Gorilla glass is used for the display window rather than cheaper plastic :) It is what I would normally expect to find when handling professional radio equipment used by law enforcement.... solid, dependable and built to last, even in harsh conditions. You can tell that I like the design can’t you  ;D One surprising omission is a 1/4” tripod socket or an accessory tripod mount to fit the two recessed threaded mounts that are used for a lanyard option. I shall fix that with my own tripod mount.

I will not say too much regarding the performance of the a Reveal Pro as I have yet to have some serious time using it and studying its behaviour in differing scenarios. What I can say though is that the images that it was producing during a quick test yesterday were more than acceptable and I am hard to please ! The additional resolution over the standard a Reveal is a significant enhancement and it is clear that the noise management routine is working well and is effective. The Reveal Pro apparently uses a Generation 3 imaging core. I am thinking that Seek Thermal may actually have been too ‘honest’ with this camera in that temperature spans as low as 1 Degree C are possible. In a low Delta T scene the camera appears to select a very narrow span and this inevitably brings with it a rise in image noise content. Other OEM’s often limit the minimum span to prevent the camera displaying too much noise on the display. They effectively hide the noise out of sight of the user. This is important to understand when comparing thermal cameras against a low Delta T scene. The Seek Thermal Pro core will have a higher noise output than the larger pixel equipped microbolometers but I understand the physics of the situation so will not be overly critical, especially as the Reveal Pro offers such small temperature spans that are normally only needed for specialist investigations. At more common temperature spans the noise content of the images is more than acceptable, as I have already stated. Image definition may be suffering a little compared to a larger pixel microbolomter but I will need to study this more closely.

I will now get to know the Reveal Pro and do some tests on its imaging performance. I am under no illusions that some image processing ‘magic’ is working very hard inside this camera to tidy up the picture. The question is, does that image processing get in the way of producing decent imagery ? I suspect I will be pleasantly surprised.

For information, I work in monochrome and Iron palettes when thermal imaging in most scenarios. High contrast rainbow palettes can be very unflattering to a thermal camera, especially where noise content is concerned !

Finally, am I happy with what my £440 purchased ? Most definitely !

My thanks to Bill W and Ultrapurple for bringing the Reveal series to my attention  :-+

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 21, 2020, 05:37:13 pm
I have also purchased a case for my Reveal camera. Not the overpriced Seek offerings, but one of the many generic camera case offerings at more sensible prices ! It was only £10 but may be a tight fit so hold off buying one until I receive and test it. The case is a horizontal type with strap or belt mounting :)

https://www.ebay.co.uk/itm/Vanguard-BIIN-II-7H-Compact-Camera-Case-Black/401559868700?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649 (https://www.ebay.co.uk/itm/Vanguard-BIIN-II-7H-Compact-Camera-Case-Black/401559868700?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649)

https://www.cameracentreuk.com/vanguard-biin-ii-7h-compact-camera-case-black (https://www.cameracentreuk.com/vanguard-biin-ii-7h-compact-camera-case-black)

It is also available from Amazon

https://www.amazon.co.uk/BIIN-II-7H-BK-Horizontal/dp/B01CUL1EN8?th=1 (https://www.amazon.co.uk/BIIN-II-7H-BK-Horizontal/dp/B01CUL1EN8?th=1)

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: triplex on March 21, 2020, 11:02:31 pm
Do you guys know if the lack of video recording is a hardware limitation or they just didn't bother to write firmware for it?
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 21, 2020, 11:31:13 pm
Sadly only Seek Thermal can answer that question definitively.

Let us consider the hardware though....

The Seek Reveal standard model running at <9fps does not get very warn in use. The Reveal Pro 15fps does get warm in use. The Reveal has a run time of 10 hours versus the 4 hours of the Reveal Pro. It would be fair to say that the higher resolution Pro model requires more computing power to produce its improved images. This extra computing power could be just for the image acquisition and display, but I suspect it is also the more advanced image processing that the Reveal Pro appears to have. If the camera were to capture, display and record the image data at 9fps or 15fps, it would increase the load on the processor. That could potentially lead to significantly lower run time and additional heat generation that must be dissipated. There is a lot going on in a small casing in the Reveal so heat could be a problem. Too much and the performance of the imaging core could be degraded. There is also the possibility of an overheating processor that then lowers the frame rate to cool itself. Users might also complain of the casing becoming hot !

Those are the hardware arguments but it is possible that SEEK Thermal decided not to equip any Reveal model with a video recording mode. I expected to see such a mode on the Reveal ShieldPro as that is intended for law enforcement. The same applies to the Reveal FirePro. Video recording is a useful feature for both law enforcement and fire fighters during training. Then there is the lack of any wireless data capability in spite of such clearly being planned in the design stages as the PCB has pads for the components. A bit of a mystery that one  :-//

So why would Seek decide to not equip the Reveal with a video recording function ? Well the marketing team may have plans for other cameras that would offer video recording and do not want the Reveal to impact the potential sales of the new product. But since 2016, when the Reveal Pro was released, we have not seen a similar format “Reveal PLUS” model that offers video recording or wireless connectivity. If Seek had really wanted to provide video recording, then they would have found a way. It could just be that the effort to achieve such in the Reveal was just too great to justify the cost. A more powerful processor would likely generate more heat and a redesign of the heat dissipation paths would have been required.

So not a definitive answer, but hypothesis. Only Seek have the definitive answer as a I said at the beginning.

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 21, 2020, 11:39:12 pm
Another question that, for me, is the “Elephant in the room” is why the Reveal series only save a standard image file to flash memory and not a fully radiometric image file for further analysis on a PC ? Also, why is it not possible to stream video from the Reveal USB port like a web cam. Both features would have been seriously useful to many users.

I have no answers to either question. An opportunity missed though.

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 22, 2020, 12:09:49 am
Looking at the datasheet for the Seek Shot it can be seen that a compact camera can be capable of both video recording and streaming of the image to another device. We must consider the fact that the Seek Shot is <9fps but even a video recording mode that dropped the Fast Frame camera down to 9fps to assist the processor in its task would often be acceptable to a user.

https://www.thermal.com/uploads/1/0/1/3/101388544/seek_shotpro_specsheet-1.pdf (https://www.thermal.com/uploads/1/0/1/3/101388544/seek_shotpro_specsheet-1.pdf)

It would seem that the lack of video recording, radiometric image files and streaming capability on the Reveal Pro was a conscious marketing decision by Seek Thermal. They wanted the seek Reveal Pro to be a “grab and go” device for professional users but without the high price tag associated with their competitors products. Adding hardware like Wi-Fi and a PC program to analyse saved image data would have added cost to the project. The absence of Video Recording remains a mystery as it is provided on the cheaper Seek Thermal dongle cameras. I return to my thoughts on processor loading and heat production though. Some users of mobile phones and dongle cameras have reported their phones getting very hot and processor thermal speed throttling when recording video. Video recording is clearly a significant additional load.

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 22, 2020, 12:01:20 pm
I raised a post asking about supplemental lens adapters for the Reveal series.

https://www.eevblog.com/forum/thermal-imaging/seek-reveal-and-reveal-pro-supplemental-lens-adapters/ (https://www.eevblog.com/forum/thermal-imaging/seek-reveal-and-reveal-pro-supplemental-lens-adapters/)

This lead me to an eBay seller called LogoXXL (Wolf) And I invited him to join the forum. He has a lot of Germanium lenses and has designed replacement lenses for thermal camera. One design is fir the Seek Reveal. After joining the forum today, he has kindly uploaded  derails of modifying a Seek Reveal and adding a better lens. He has included examples of the pictures produced and comparisons with the original and new lens images. It is clear from the pictures that the Seek Reveal lens is a weak component in the camera design. I suppose you could call it ‘good enough’ rather than good.

Having spoken to a Bill W regarding Chalcogenide IR Glass lens design it is apparent that it is an art and the experience of the designer greatly influences the quality of the optical system and images produced. The Seek lens may be just a bit too small or not optimised for best image. I would look at fitting a better lens if I could find one that would match the Reveals current lens mount. Sadly no third party replacement lenses are available in the same format. Note also that the example images from LogoXXL using the 35mm lens in place of the Seek  lens  did not show obvious lens characteristic compensation errors which is good  :-+

Fraser

EDIT: PDF was corrupted by my iPAD  :palm: Now uploaded from my laptop and checked. Sorry.
I note that it exceeds the maximum file size that was allowed, yet it has uploaded fine.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on March 23, 2020, 01:06:09 am
Isn't a significant part of the difference between the clarity of the original Seek images and LogoXXL's telephoto images due to the fact that the stock lens is "fixed focus" and "fixed focus" == never focused?
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on March 23, 2020, 09:09:10 am
Quote from: Fraser on March 21, 2020, 11:39:12 pm
Another question that, for me, is the “Elephant in the room” is why the Reveal series only save a standard image file to flash memory and not a fully radiometric image file for further analysis on a PC ? ...

Fraser

???

Quote from: Ultrapurple on February 28, 2020, 08:19:57 pm

Images are normally stored as non-radiometric PNG. A menu option lets you save all temperature data, in which case the image is stored as a radiometric TIFF.

 ;D  ;D
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 23, 2020, 09:45:22 am
IwuzBornanerd,

Interesting ! I had read the review but forgot about the TIFF comment.

Reading the Reveal manual suggested that the Reveal saves files in PNG format with only the spot temperature contained in the file. I can see I will have to do more reading and see exactly what is available in the way of radiometric data. There does not appear to be a SEEK Thermal PC program for image analysis though  :-//

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on March 23, 2020, 09:49:11 am
Quote from: Fraser on March 23, 2020, 09:45:22 am
There does not appear to be a SEEK Thermal PC program for image analysis though  :-//

Fraser

No, but I have a crude one--for Linux at least.  Well...maybe not a lot of analysis though; whatever that entails.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 23, 2020, 10:14:57 am
This raises a good point about the Seek Reveal Pro..........


When you receive the unit, the box contains a piece of paper that only really tells you to register the unit. There was no obvious web link to the Reveal Pro user manual (unless I missed it!) I went on the Seek Thermal site and searched for the Reveal Pro user manual. I was directed to download the Seek Reveal manual but that does not cover the Pro model differences ! A google search for the Reveal Pro found a simple guide that could not truly be considered a user manual yet that is its title. See here......

https://res.cloudinary.com/iwh/image/upload/q_auto,g_center/assets/1/26/Seek_Thermal_RevealPRO_Handheld_Thermal_Imaging_Camera_User_Manual.pdf (https://res.cloudinary.com/iwh/image/upload/q_auto,g_center/assets/1/26/Seek_Thermal_RevealPRO_Handheld_Thermal_Imaging_Camera_User_Manual.pdf)

Now in comparison, I can search for Reveal ShieldPro or Reveal FirePro and I find proper user manuals.

Am I missing something here or is the Seek Thermal documentation for the a Reveal sadly lacking  :-//. Granted I have not spent a lot of time searching but this particular model seems to be treated a bit like an afterthought by Seek Thermal. I have yet to find any useful guidance on the radiometric ‘full’ mode and any associated Radiometric TIFF image storage. I would have expected Seek to highlight this units radiometric capabilities. Saying that, the specification for accuracy is stated as +/-5C or 5%, whichever is greater.

Maybe I am just being dumb here (there is a lot going on at the moment) but I expected to find a proper Reveal Pro manual and not a generic “Reveal Series” manual.

Seek Reveal ShieldPro manual:

https://www.thermal.com/uploads/1/0/1/3/101388544/reveal_shiledpro_user_manual.pdf (https://www.thermal.com/uploads/1/0/1/3/101388544/reveal_shiledpro_user_manual.pdf)

Seek Reveal FirePro manual:

https://www.thermal.com/uploads/1/0/1/3/101388544/reveal_firepro_manual_-_jun19mn.pdf (https://www.thermal.com/uploads/1/0/1/3/101388544/reveal_firepro_manual_-_jun19mn.pdf)

Seek Reveal “Series” manual

https://static1.squarespace.com/static/54d4e995e4b08c8065ddef4c/t/570e7e7ff699bb1e4671e3dc/1460567680755/Reveal-English.pdf (https://static1.squarespace.com/static/54d4e995e4b08c8065ddef4c/t/570e7e7ff699bb1e4671e3dc/1460567680755/Reveal-English.pdf)

And a “User Manual” for the Reveal like the Reveal Pro that is more like a QSG !

https://www.thermal.com/uploads/1/0/1/3/101388544/reveal_series_user_manual_051117.pdf (https://www.thermal.com/uploads/1/0/1/3/101388544/reveal_series_user_manual_051117.pdf)

The official Seek Thermal web site page for Seek Reveal Series user manual ...... pretty rubbish IMHO. No mention of the Pro, Shield or Fire models.

https://support.thermal.com/hc/en-us/articles/207894336-Seek-Reveal-User-Guides (https://support.thermal.com/hc/en-us/articles/207894336-Seek-Reveal-User-Guides)


The Seek Q&A section answers common questions but offers no comment on the Pro TIFF files.......

“Q: How do I download/copy images?

A: To download an image, remove the rubberized back door from the end of the device and insert your included USB cord from your computer into the TIC. If you are using a Windows or MacBook, you will be automatically prompted to import all your photos. The photos are rendered in a standard .PNG format and can be edited, annotated, and inserted into documents easily. If you have a Reveal or RevealXR unit, you can eject the 512mb/4GB SD card, and erase and reuse it with an adapter (not included). When placing the SD card back in notch, make sure and the contacts are facing downward, and try not to bend the card or force it into the slot.”

To be honest, I find the documentation to be pretty awful ! Users should not have to hunt for manuals, hunt for information on their specific version or fail to find a detailed description of their versions capabilities and how to use them.

Ultrapurple ..... any thoughts on this ? Am I just being dumb here ?  :-//

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 23, 2020, 12:12:53 pm
OK, after some searching I see that our very own Joe_C has built Seek Reveal Pro TIFF file support into his ThermoVision software :-+ What wonderful chap Joe is  :-+ :-+ :-+

http://joe-c.de/pages/posts/unterstuetze_kameras_supported_cameras_195.php (http://joe-c.de/pages/posts/unterstuetze_kameras_supported_cameras_195.php)

I shall have to try it out on some test images  :)

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 23, 2020, 12:52:12 pm
It seems Joe_C was having the same problem as me finding data on the Reveal Pro radiometric TIFF file format !

From his 2018 update ......

“ - Seek Reveal Pro tiff frames can be read... a poor implementation now... i only grab the frame which is remapped with given min and max values or with a 2 point calibration with slope and offset. sadly i know not enough about the format, and seek seems to tell nothing about...”

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 26, 2020, 04:11:55 pm
I have found an excellent padded case for your lovely little Seek Reveal series camera. It could have been custom made for the role as the fit is perfect. The case may be worn on a belt or the provided shoulder strap attached. I wanted a horizontal belt format and this is perfect. It has both a Velcro and a zip closure and the zip can be left open fir fast access it closed for security :) The case is the Vanguard BIIN 7H and cost me £10.39. They have returned to their normal price of £12.99 on eBay now and £13.99 on Amazon.co.uk. I attach pictures of the nice fit :)

 https://www.ebay.co.uk/itm/Vanguard-BIIN-II-7H-Compact-Camera-Case-Black/401559868700?_trkparms=aid%3D111001%26algo%3DREC.SEED%26ao%3D1%26asc%3D225076%26meid%3D28be82eaf59144a4a960b74c4d17ecba%26pid%3D100675%26rk%3D2%26rkt%3D15%26mehot%3Dnone%26sd%3D274314951216%26itm%3D401559868700%26pmt%3D0%26noa%3D1%26pg%3D2380057&_trksid=p2380057.c100675.m4236&_trkparms=pageci%3A038c85d6-6f7c-11ea-b58f-74dbd1803803%7Cparentrq%3A179a028c1710aa66c9b2aaaafffe9334%7Ciid%3A1 (https://www.ebay.co.uk/itm/Vanguard-BIIN-II-7H-Compact-Camera-Case-Black/401559868700?_trkparms=aid%3D111001%26algo%3DREC.SEED%26ao%3D1%26asc%3D225076%26meid%3D28be82eaf59144a4a960b74c4d17ecba%26pid%3D100675%26rk%3D2%26rkt%3D15%26mehot%3Dnone%26sd%3D274314951216%26itm%3D401559868700%26pmt%3D0%26noa%3D1%26pg%3D2380057&_trksid=p2380057.c100675.m4236&_trkparms=pageci%3A038c85d6-6f7c-11ea-b58f-74dbd1803803%7Cparentrq%3A179a028c1710aa66c9b2aaaafffe9334%7Ciid%3A1)


https://www.amazon.co.uk/BIIN-II-7H-BK-Horizontal/dp/B01CUL1EN8/ref=sr_1_2?dchild=1&keywords=Vanguard+biin+7h&qid=1585238985&sr=8-2 (https://www.amazon.co.uk/BIIN-II-7H-BK-Horizontal/dp/B01CUL1EN8/ref=sr_1_2?dchild=1&keywords=Vanguard+biin+7h&qid=1585238985&sr=8-2)
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 27, 2020, 03:46:30 pm
Another case for the Seek Reveal that does not cost a fortune and fits well. It is a vertical model for those who prefer such orientation.

The Jessops (UK) “Trek2 Compact Case” Medium sized case with both zip and Velcro fastening. It has both belt and shoulder strap mounting options. This one cost me £2.50 + £2.90 postage. It is well made and offers the Reveal excellent protection.

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on March 31, 2020, 06:08:03 pm
I finally decided to jump into the thermal image camera world and purchased a Seek Reveal PRO FF. I selected this camera based on: reasonable thermal resolution of 320X240, all-in one unit vs a dongle, good battery life of 4 to 5 Hrs, solid construction (see mikeselectricstuff teardown video), and reasonable price ($608  at Global Test Supply). The $608 included a Reed R8500 Video Inspection Camera that Global Test Supply threw in.

I won't do a review of the Seek Reveal since Ultrapurple has done a pretty comprehensive one at the start of this thread. Also there are other reviews on Youtube.

The documentation that Seek provides with the camera and on-line doesn't say it, but you can get radiometric data from the camera (sort of). On the Reveal, if you select Normal in the Modes menu the camera saves a PNG file. If you select Full Frame in the Modes menu and select Thermal Data "On" in the Settings -> Image -> Thermal Data menu, the camera saves a PNG file and a 3-page TIFF file. The second page of the TIFF file has 320X240 floating point values. As IwazBornanerd guessed, these are temperature values. The temperature values are always in degrees C no matter what you have selected for temperature units on the Reveal (C,F or K). The first page of the TIFF is an ARGB RGBA image, and I think the third page is raw thermal values since the data in the third page is 16-bit integers. As far as thermal metadata is concerned the MakerNotes has e=0.97, tr=25 in ASCII. Must be emissivity and reflected temperature. You can select emissivity on the Reveal (four values). I didn't see a place to change reflected temperature. This is a lot different than the thermal metadata saved in a FLIR file where they save many dozens of metadata tags.

The second page of the TIFF file is of interest from a radiometric data standpoint. You could map the temperatures to a different palette than is used in the PNG file, or map the palette over different temperature ranges, or show temperatures at points, or on a line, or in an area as is done with FLIR tools and other software for FLIR files. However, that's about it. You couldn't change the emissivity for example because you don't know how temperatures are calculated vs emissivity. That's why I say above that you "sort of" get radiometric data in the TIFF file.

A couple of oddities about the Reveal TIFF file --- the image in the first page seems to use a different palette than the PNG file. Not sure why. Also the images are rotated 90 degrees CW. In order to more easily work with the second page of the TIFF file, you could first split the original TIFF into three separate files. You can do that with ImageMagick convert.

So as an example, I took the requisite picture of my hand for the image posted below. In order to view the temperatures in readable form I wrote a little program in CodeRunner (on macOS) to create a csv file from the temperatures. The csv file also posted below, uses a comma as a separator and \n for line endings. Also added column headings 0...319 and row headings 0...239. The program also maps a simple 256 entry grayscale palette over the minimum to maximum temperatures and saves a ppm (P6) image which is also posted below.

The minimum and maximum temperatures from the csv file are 16.94 C (62.5 F) and 28.57 C (83.4 F) respectively. In another thread, Fraser posted about the temperature accuracy of the Seek Reveal. Basically the accuracy is specified as +/- 5 deg C or +/- 9 deg F. That looks to be about what it is. The background in the picture is an inside wall where the ambient temperature is 70 deg F. Also, skin temperature on the hand is about 5 deg F lower than body core temperature. It would be nice to get a little more accuracy on the temps.

Jim



Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 31, 2020, 07:32:14 pm
Very interesting. Thank you  :-+

I have not had time to play with the TIFF file yet.

I found the Reveal Pro documentation very poor and the lack of a software application to read the Radiometric TIFF files disappointing. I expected a PC and MAC program that permitted adding temperature measurement points, selecting span/centre temp for the displayed image and a change of Palette as a minimum.

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on March 31, 2020, 07:35:37 pm
A contact in the industry has advised me that Seek only sell their cores with a lens fitted. They wish to ensure that the user gets the best experience possible with their core. As such Seek Thermal do not offer alternative lenses for use with their cores or cameras.

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on March 31, 2020, 09:36:21 pm
Good work JimM.  :-+ Thanks for the tiff.  Please see the following thread:

https://www.eevblog.com/forum/thermal-imaging/seek-compact-pro-radiometric-jpegs/?PHPSESSID=oq1rsl36so8m87o6givc6h6kd0 (https://www.eevblog.com/forum/thermal-imaging/seek-compact-pro-radiometric-jpegs/?PHPSESSID=oq1rsl36so8m87o6givc6h6kd0)

As you surmise, there are some nifty things you can do with the temperature list.  Such as change palettes and extract more detail (and show all the noise).  I offer some samples.

[attach=1]

[attach=2]

[attach=3]

If you are so inclined you can fairly easily write a routine using libtiff to extract the 2 data frames.  When I did that, however, I got the third frame with only 4 non-zero numbers!  Maybe you have better luck.  I have not had that happen with any other tiffs I've used it on though.

I can believe that the 3rd frame is representative of radiance above -40C, but the values are a far cry from anything I would call "raw".  I'll attach what I got from your tiff via libtiff...even the worthless frame 3.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on March 31, 2020, 10:47:24 pm
IwuzBornanerd,

Yeah, you're right -- there are only four non-zero values in the third frame (or page).
I actually didn't look at that page for the image I posted.
I was thinking it was raw data based on a previous TIFF someone posted for a Seek compact PRO camera.
When I opened the third page of that file in GraphicConverter (Mac app) it showed a Grayscale image.
GraphicConverter typically does a contrast stretch of low contrast images like you get with raw data from an IR camera so you can actually see something.
So now I don't know what the third page represents.

It would be interesting to find out what other people get for that page from a Seek Reveal.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on April 02, 2020, 04:37:02 am
To do more testing of the Seek Reveal PRO, I took a picture of the inside of the refrigerator where there are negative Celsius temperatures. The temperature range shown on the PNG image seems to be correct. However all the negative temperatures in the TIFF file temperature page are 0x80000000 (little Endian). The C function fprintf() interprets that as 0. So apparently there is a software bug with negative temps (or else I have a defective camera). The camera has software version 1.0.6.9 which is supposedly the latest software. So again, it would be interesting to see what others people get for a TIFF file.

Edit: I take that back about the negative temps being wrong in the TIFF file. They are correct. For some reason, when I split the file in ImageMagick, I get the bogus values. If I manually excise the second page of the TIFF file, I get the correct temps. Minimum -16.72 and maximum 7.25 C respectively. The PNG and TIFF files are attached.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on April 02, 2020, 09:42:26 am
Still only 4 values in page 3; bummer.  Different values though.  It will be interesting to see if this is consistent across all Reveals or if there is something unique in your unit, as perhaps software/firmware version.

The beauty of these tiff files is that regardless of how wide the temperature span is you have the temperature value from each pixel out to a ridiculous 6 decimal places so you can re-create the image with more contrast & detail if desired.

With the same palette spread over a narrower temperature range:
[attach=1]

Or with a more contrasty palette, which might not make it any easier to identify things (I need to make something better than the 2 I have):
[attach=2]

Or you can just apply another standard palette  :):
[attach=3]
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on April 02, 2020, 12:31:18 pm
I will upload my Reveal Pro TIFF file when I find a moment today.

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on April 02, 2020, 02:34:33 pm
Yeah, the 4 non-zero values on page 3 seems to be a constant across all images. Not the exact same 4 values, but just 4 values. I also wonder if that's unique to this camera or not. I have software version 1.0.6.9, which according to the Seek website is the latest software.

Regarding the 4-byte floating point values, yes half-precision 16-bit point floating point would be perfectly sufficient considering the accuracy of 5 deg C. However, 16-bit floating point although used in image processing, isn't a standard data type in C for example.

Regarding the image itself, I was wondering what that scalloping affect at the top of the image was, since there is nothing with that shape on that shelf. There is an open egg carton on the shelf below the tupperware container and I think that's a reflection of the egg carton. I have seen reflections like that in numerous thermal images -- reflections that you don't have in the visible light world, or are greatly attenuated. Interesting.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on April 02, 2020, 03:27:37 pm
Pictures from my Seek Reveal Pro FF

I had to change the file ID to .jpg but just rename it .TIFF after saving.

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on April 02, 2020, 07:12:49 pm
Fraser,
Ok it looks like you have the same version of camera software that I have and you get similar results.
For your IMGT0056.TIFF file 2nd page, I get 11.8 to 30.2 C temps.
For the 3rd page there are only 2 4 non-zero values!

So I guess our cameras are working the same.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on April 02, 2020, 09:26:16 pm
I get 4 values from both Fraser's images--the same values even AND the same values that are in JimM's image 20!  And upon closer examination the 2 that are different in Jim's file are higher by 16 compared to the other files.  Strange.

I should have known there would be a cat photo.  :)  But I thought it there might be something more "technical".

Standard palette, narrower span:
[attach=1]

[attach=2]

"Contrasty" palette:
[attach=3]

[attach=4]

I hope that's enough for now (wanted to get this in before Fraser retires for the day).  I'll attach the tiff info. below...IF I don't exceed size limits.

Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on April 02, 2020, 09:46:14 pm
Bornanerd,

Ah, you're right, its 4 non-zero values on page 3 -- it's 16-bit integers, not 32-bit integers.
Maybe page 3 is work in progress -- not ready for prime time.

Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on April 02, 2020, 10:56:40 pm
IwuzBornaNerd,

In my defence, those were some pictures I took on the day my Reveal Pro arrived  ;D

Our cats are ever present thermal ‘targets’ and their fur can act as a pretty good detail test  :-+

I can produce some test target images if you really want boring old inanimate targets  ;D

Thank you for analysing the images  :-+

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on April 03, 2020, 12:13:08 am
Quote from: JimM on April 02, 2020, 09:46:14 pm
Bornanerd,

Ah, you're right, its 4 non-zero values on page 3 -- it's 16-bit integers, not 32-bit integers.
Maybe page 3 is work in progress -- not ready for prime time.

Well, page 3 is fully populated in images from the Android app for the "Compact" dongles, as you can see on the Seek Compact Pro Radiometric JPEGs thread.  Maybe for some reason not on Reveal, but why...

How much of the on-device storage space can you access?  Just the image directories or is there something else you can see?  I'm wondering if it's possible to get at the calibration files like the ones that the Compacts spit out on startup.  An equivalent to the startup frame 9 would be interesting.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on April 03, 2020, 12:21:19 am
Quote from: Fraser on April 02, 2020, 10:56:40 pm
IwuzBornaNerd,

In my defence, those were some pictures I took on the day my Reveal Pro arrived  ;D

Our cats are ever present thermal ‘targets’ and their fur can act as a pretty good detail test  :-+

I can produce some test target images if you really want boring old inanimate targets  ;D

Thank you for analysing the images  :-+

Fraser

You're welcome.  :)   Test targets would be most interesting IF we had values in page 3, but I'm not sure they would be all that useful otherwise.  It might be instructive, though, to see what happens as you increase the temperature of a black body up towards 330C.  I would expect, based on the saturation that I have now seen, that some pixels would hit a limit before others and maybe some would get all the way to 330C while others do not.  So if you image a uniform temperature surface at or near 330C, the pixel temperatures would be "all over the map".
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on April 03, 2020, 01:18:00 am
Quote
How much of the on-device storage space can you access?

Here's what I see:

The DCIM directory has all the images: PNG and TIFF
DEVICE.TXT has a few lines about software version, etc
The other directories have from 10 to 20 files each.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: hinofi on April 04, 2020, 11:35:36 am
Quote from: Ultrapurple on February 29, 2020, 09:23:00 pm

Results from FLIR SC-660:

(Attachment Link)

Results from Therm-App Pro:

(Attachment Link)


Results from Therm-App:

(Attachment Link)


Results from SEEK Reveal Pro:

(Attachment Link)


Good work, but I think you should compare them at same temperature span, not only same scene.
BTW, Flir's high-end camera still have that 4 C degree minimum span limit? Rediculous.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on April 05, 2020, 02:04:11 am
Quote from: JimM on April 03, 2020, 01:18:00 am
Quote
How much of the on-device storage space can you access?

Here's what I see:

The DCIM directory has all the images: PNG and TIFF
DEVICE.TXT has a few lines about software version, etc
The other directories have from 10 to 20 files each.

Thanks for the listing.  It looks like all operating system stuff to me, but they could certainly put their stuff in with the regular system stuff.  I won't ask you to sift through it all since I can't say exactly what to look for.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on April 05, 2020, 03:36:26 am
So, now that I've had the SeeK Reveal PRO camera in my grubby paws for all of six days, it's time for some comments and criticisms of the unit. Generally overall, I'm happy with the camera and think it's pretty good value for the  money, but constructive criticisms are always in order. Listed below are comments -- some are nits, while others are more substantive.

That's all the comments I have for now. Sure, it would be nice to have video, but the camera already gets noticeably warm after about 1/2 hour of use (not uncomfortably so, but it does get warm), so I can understand why video wasn't included.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on April 05, 2020, 11:29:48 am
Some comment on your comments  ;D

1. Yes the documentation is sadly lacking for the Reveal Pro. I wish they had provided a URL to a decent manual that reflected the ‘PRO’ status that Seek claim for the Reveal PRO model :(

2. It is common for USB chargers to not be supplied with equipment these days. Mobile phone manufacturers were the first to do this. The ideology behind the move is that most buyers already have one or more USB chargers and that may be used. The idea is to reduce waste and duplication. Not convenient for those without a suitable USB charger but, as you say, such chargers are common and cheap. HOWEVER from a manufacturers standpoint I would be worried about nasty poorly designed chargers damaging My products charge circuit so would want to provide a decent charger to prevent warranty claims.

3. Power on is a quick press of the central button. Power off is a long press of the central button. The camera then goes into shutdown with a “Please wait” display as it gets its affairs in order (managed shutdown of Linux)

4. The unit has the PCB tracks and pads for Wi-Fi but sadly no model released to date has the parts installed :( I agree that this omission is a pity.

5. I remember reading “somewhere” in one of the reveal manuals that you can delete files in the Reveal from your PC without issue. I strongly suggest you do not use a format command though as I am unsure of the Reveal Pro flash memory architecture. Formatting a FLIR E4 picture memory has caused issues. The standard Reveal cameras use a removable micro SD Flash card and the camera just acts like a USB card reader in PC Mode so a format would be permissible but pretty pointless anyway. Just select unwanted picture files and delete them using the camera option or via the PC connection.

6. Accuracy is a bit of an issue with small cores like the FLIR LEPTON and Seek Thermal offerings. The manufacturers are being honest about the accuracy for which I congratulate and respect them. These compact cores are built down to a price and compromises were made. I wrote a piece on those compromises that I will see if I can find in my emails. Basically, the requirements for accurate temperature measurement are not met by these tiny cores so accuracy suffers. For many tasks this can be tolerated as a downside of the lower cost and affordable product. The temperature accuracy of the Seek Core has been known to improve after a temperature stabilisation period of up to 5 minutes. I will be testing my Seek Reveal Pro measurement accuracy against professional blackbody sources soon. The manufacturer has chosen to be honest and declare relatively poor accuracy but that does not mean the core exhibits such a wide tolerance after a temperature settling period and in reasonable ambient temperatures. We shall see. We also need to consider the preset emmissivities rather than user set precise emissivity for a target. Such could also influence the accuracy specification provided by Seek.

Fraser

Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on April 05, 2020, 12:20:20 pm
I just found my comments on the accuracy specifications of the Seek cores. It was actually a post on this forum....

https://www.eevblog.com/forum/thermal-imaging/seek-thermal-core-measurement-accuracy-whats-the-story/ (https://www.eevblog.com/forum/thermal-imaging/seek-thermal-core-measurement-accuracy-whats-the-story/)

Fraser
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on May 02, 2020, 09:25:08 am
I discovered yesterday that it was very simple to grab the palettes off Seek app images for use in my manipulator program so I did that & then applied them to some of the tiff file temperature lists to see how well they matched the color image stored in the first frame/page of the tiff file.  But when I opened the Reveal Pro tiff files with any of the image viewers on my Linux box the colors didn't match ANY of the palettes.  It looks like those images have the Red & Blue swapped.  This is NOT true of any tiff's I have seen off of the Seek app--either those posted by others on this forum or those I have gotten off my cameras with the Seek app.  Do the Reveals use BGR while the app uses RGB??

I present a collage of samples below.  The upper left image is from a tiff uploaded by a forum member & the lower right is part of the image my program got from the temperature list in that tiff file & it is a good match for the upper left one.  The rest of the images are Fraser's & JimM's tiffs displayed by ImageMagick.

When I display the tiff temperatures in my program using the Seek palettes I get very good match with the tiff color images, but not with the Reveal tiffs.

Do the tiff's come out that way for you guys?
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on May 02, 2020, 08:23:35 pm
Quote
Do the tiff's come out that way for you guys?

Yeah, I noticed that. To further your guess that maybe they just mixed up the colors in the tiff file palette, I took a picture of my coffee cup with the "White" grayscale palette (white is hot). In that case  r=g=b for each palette entry. So it wouldn't matter if they mixed up the color order. And yes, with that palette the tiff image and png image are the identical (IMGT0029 files below). When I switched to the "Iron" palette I get IMGT0030 files below. (Better drink that coffee, it seems to be cooling off a lot).
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: IwuzBornanerd on May 02, 2020, 09:48:21 pm
Thanks for the feedback.  Glad it's not just me, but then you are on a Mac & I'm on Linux; I have to wonder if Windows sees it the same way.

If you open the tiff in GIMP & from the Colors menu select Components> Channel Mixer you can set the red channel to 0 red & 100 blue and the blue channel to 0 blue & 100 red and get the same color image as your png:

[attach=1]

Which matches well with what I get from the temperature list using the Seek Iron palette:

[attach=2]
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on May 03, 2020, 12:05:08 am
Just to close the loop on this, I took the first page of the tiff file and swapped the Red and Blue channels in ImageMagick. The resulting image is then identical to the png image. So, yes the tiff image is bgr rather than rgb. For reference, the ImageMagick code is:
Code: [Select]
convert input.tiff -color-matrix '0 0 1 0 1 0 1 0 0' output.tiff
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on May 03, 2020, 12:14:01 am
An interesting development chaps. Well done  :-+ A firmware bug maybe ? I wonder if Seek already know about it. It might be worth an email to them ?
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: JimM on May 03, 2020, 07:21:11 pm
I would call it a bug. However, I noticed another thing that I think is also a bug. If you go into Gallery mode on the Reveal and delete an image, if that image has an associated tiff file, you can't go back and delete other files in Gallery, or sometimes the screen just goes black. If you then connect to your computer and look at the image directory, you will see that only the png file was deleted. The tiff file is still there. So, delete the tiff file on your computer, eject the Reveal volume, disconnect from the computer (which turns off the Reveal camera), and then turn the Reveal back on and go into Gallery mode and you will see that Gallery works again. By the way, I think having the camera turn off when you disconnect the USB port should be changed. Why have the camera turn off at that point? So there are several things that could be changed in the software/firmware on the Reveal.
Title: Re: SEEK Reveal Pro FF review by Ultrapurple
Post by: Fraser on May 03, 2020, 08:14:46 pm
It is like the Reveal’s Radiometric TIFF function is either an afterthought, or is known to be buggy. I did not find any great mention of the Reveal PRO radiometric TIFF mode in documentation and SEEK have not provided a file viewer for the radiometric TIFF images.

Fraser