Seek Scan

[triplex]

https://www.thermal.com/seekscan.html
https://www.thermal.com/uploads/1/0/1/3/101388544/seek_scan_specification_sheet_mar20.pdf
New device from a well known (to us) brand.

Thermal Sensor Resolution 206 (h) x 156 (v)
Sensor Sensitivity 40 mK (typical), <50 mK (max) @ 25°C (Post Signal Processing)
Frame Rate <9Hz
System Requirements Supported Operating Systems: Windows 7 and 10

As far is i know first seek device with native windows software. I wonder if this is a new sensor  for increased sensitivity or tweaked design of previous devices, other seek cameras have documented sensitivity of 70mK.

[Fraser]

Scan measurement time is 1 second. OK for static checks but inadequate for crowd or moving target checks.

The long ‘scan’ time of 1 Second at <9fps suggests plenty of post image capture processing, likely stacking images to achieve the claimed <50mK NeTD.

I would be interested to know the construction of their temperature reference emitter. The system likely uses “lock-in” thermography with the temperature reference to achieve the uncharacteristic (for Seek Cores) temperature accuracy of this system.

Fraser

[IwuzBornanerd]

The "sensor sensitivity" spec does state "Post Signal Processing", but the spec's also include F/1.2 lens instead of the usual F/1.4 used in the "Compact" versions (unless that spec refers to the visible camera).  I would expect it also is biased for a narrower sensing range, but maybe the stacking or other "processing" was deemed adequate so no hardware changes were made,

I expect the software is so specialized that even if it will drive the Compacts it would not be useful for other purposes.  Now that they have something for Windows, though, they might be more likely to offer something for the Compacts that runs on Windows.

[Fraser]

The thermal sensitivity of a camera is an area where specifications can get very ‘muddled’ by the marketing department 

From what I understand from Bill_W NeTD is supposed to be specified with an F1.0 lens used in the test. Any slower lens subsequently fitted to the system will degrade the sensitivity in line with the F number.

“Sensor” thermal sensitivity is a term commonly used to indicate what the microbolometer is capable of with a certain lens fitted. As stated, with NeTD, this is normally an F1.0 lens. If not stated in the specification as NeTD, any lens specification could be involved, even an F0.7 !

I may be being unfair to Seek though and they have just kept their specifications simple for the intended readership.

Fraser

[Vipitis]

regarding the lens:

form the website there is a background video embedded, the direct url is here: https://www.thermal.com/uploads/b/101388544-385591210749727708/seek_scan_product_page6_262.mp4 jump to 0:25 and get a detailed view of the lens from the back. the rear aperture looks tiny to me.

attached: a screeshot

[Fraser]

Vipitis,

I think that picture is actually showing the front of two different lenses and not the rear of one. Presumably intended to show that a larger, smaller FOV lens is available if required ? No lens options are mentioned though so it is likely just generic marketing video for the Mosaic core.

The larger Seek Thermal lens is that used on the Seek ShieldPro and provides a 20 Degree FOV on that 320 x 240 pixel platform. It provides a smaller FOV on the lower resolution Seek core.

The datasheet for the Seek Scan system shows the standard ‘small’ lens is normally supplied and this offers around 35 Degrees FOV.

Looking at the Mosaic data sheet we can see that the system uses a standard 4mm F1.2 lens as to provide the 35 Degrees HFOV. The other lens shown in your screen capture is the 9.2mm lens that would provide a faster F1.0 aperture and 15 Degree HFOV.

https://www.thermal.com/uploads/1/0/1/3/101388544/mosaic_core_specification_sheet.pdf

Fraser

[Fraser]

The rear element area of a Seek Thermal Mosaic lens is actually quite large as shown in these images stolen from the internet.

[IwuzBornanerd]

The rear element area of a Seek Thermal Mosaic lens is actually quite large as shown in these images stolen from the internet.

And it probably needs to be every bit that large because the front surface is divergent and the lens is thick, as can be seen in my photos on the Seek Thermal lens repair thread.  ...Assuming of course that the mosaic lens is similar to the lenses used on the "Compacts".

[Ben321]

Well this is useless, as are any of those LWIR camera setups used to detect fevers on the foreheads of large groups of people. See my post here that explains why they are useless https://www.eevblog.com/forum/thermal-imaging/thermal-image-dataset-for-fever-and-human-temperature-measurement/msg2989428/#msg2989428

[Fraser]

Ben321,

There are ways to use a a LWIR thermal camera to help detect a fever but as I have said in the thread you referenced, there are issues, no doubt about it. There are specific areas of the body that exhibit a temperature that relates to core temperatures in areas of high blood flow served by large arteries that may be read using IR techniques. When I say ‘relates to’ I mean research has plotted the relationship between the measured temperature and true core temperature.

You are correct in your comments regarding the forehead but the area of the eye socket nearest the nose is a known reliable IR temperature test point for thermal imaging. The test subject must provide a clear view of the area however so spectacles must be removed. The area is also small so scanning moving targets would be a challenge and likely unreliable.

I will not make comment on the performance of the Seek Thermal solution as I have not tested it ! The designer does appear to be making an effort to achieve good measurement accuracy using the ‘lock in’ thermography technique that is well known and accurate. The inclusion of the thermal reference in view of the camera is an essential element in this accuracy improvement. The resolution of the core used is not what I would have expected. I would have expected Seek to have used their superior Pro core with its 320 x 240 pixel resolution. If you look at the example images of test subjects you will see that the Seek Scan uses face identification software to target the face of the test subject. It then uses a further software routine to target the eye socket in the area adjacent to the nose. This area is normally easily identifiable in the thermal domain as it glows brightly compared to other areas of the face.

As has been stated in various scientific papers, much was leant about Fever Detection using Thermal imaging during and after the SARS outbreak. Lessons continue to be learnt during the current Corona virus pandemic.

We will see a lot of thermal camera manufacturers producing Fever detection cameras as they adapt current models and cores to the task. The key is for them to understand the challenges of human body related thermography. If they take no notice of the challenges involved and just want to sell ‘Fever Cameras’ then the product may ‘look the part’ but be ineffective in practice.

Fraser

[ir.ukrm]

I got my seek scan instance

[Fraser]

Excellent 

There is a small ‘block’ visible on the Blackbody reference that I have seen in other pictures of the product. Up close does it appear to be a surface mounted thermistor/RTD or something else ? It seems unusual to have such a component on the emission plates visible face but maybe the whole of the rear of that plate is covered with a Peltier element ?

Fraser