EEVblog Electronics Community Forum
Products => Thermal Imaging => Topic started by: aargee on May 24, 2023, 10:15:31 pm
-
Hi all,
To test the sensor resolution, is there such a thing as a thermal resolution test card/pattern? Just like there is for video resolution tests (the ubiquitous colour bar/lines/grid pattern)
I imagine one could be done with a PCB with tracks that are heated by running current through them, just like a 3D printer heated bed, the problem would be thermal bleed into the board material. Another thought would be to use nichrome wire as a grid or matrix, or even simply two parallel wires, taught, at a small fixed spacing.
The problem might be image interpolation and processing in the firmware that would muddy the results (i.e. not a raw image).
Just a random thought/idea.
- Rob
-
The usual is a copper sheet, sprayed (thermal) black and separated 5mm - 15mm in front of another sheet or block also painted / treated thermal black.
Cut the pattern in the front sheet, and maintain a temperature difference between them.
With a fairly solid front sheet you can use Peltiers, or simply heat the rear sheet with heater mats. Obviously plastic screws & standoffs.
Bill
-
This is a topic close to my heart as I have black bodies and like to use them to test thermal imaging systems. What I do not have is a collection of test targets or ‘charts’ to place in front of the black body emitters. I was thinking of designing some charts and having them laser cut in aluminium.They would then be painted with a good Matt black paint.
With regard to finding black body emitters for such a project, I am finding many Dahua BB’s at great prices, such as £45 for a brand new unit ! And there are still plenty at that, or similar prices. The aperture is a 70mm x 70mm square and the face is perfect for adapting to hold a chart plate.
https://www.ebay.co.uk/itm/285209076849?mkcid=16&mkevt=1&mkrid=711-127632-2357-0&ssspo=M57-4NJjR0i&sssrc=2349624&ssuid=BgpdXaaLSkK&var=&widget_ver=artemis&media=COPY (https://www.ebay.co.uk/itm/285209076849?mkcid=16&mkevt=1&mkrid=711-127632-2357-0&ssspo=M57-4NJjR0i&sssrc=2349624&ssuid=BgpdXaaLSkK&var=&widget_ver=artemis&media=COPY)
Sadly the seller of my unit has increased the price to £54 now but he does invite offers so £45 might still be accepted.
I have images of commercial test charts for differing tests but not the exact slot dimensions.
I used a Bahtinov telescope focussing mask in my tests on the Infiray P2 Pro to show where acceptable focus was lost at close range and detail lost at longer range. I will attach an image from that test showing the test plate being used on one of my Mikron black bodies. In the accompanying test of the review I detail how the dimensions of the slots enable “pixels on target” tests to be carried out at varying distances.
https://www.eevblog.com/forum/thermal-imaging/review-infiray-p2-pro-thermal-camera-dongle-for-android-mobile-phones/ (https://www.eevblog.com/forum/thermal-imaging/review-infiray-p2-pro-thermal-camera-dongle-for-android-mobile-phones/)
If buying a decent mini Blackbody like the Dahua unit, you also gain a calibration check source for your camera :-+
I reviewed the Dahua units here…
https://www.eevblog.com/forum/thermal-imaging/danua-reference-blackbody-under-the-covers-)/ (https://www.eevblog.com/forum/thermal-imaging/danua-reference-blackbody-under-the-covers-)/)
Fraser
-
A picture of the USAF thermal test chart is attached. It looks similar to some standard optical test charts.
A test chart may also be created using materials of differing Emissivity rather than having to generate heat. Silver foil shapes on a Matt black painted board or metal plate will work. Long range test targets sometimes use the outline of a human of correct 1:1 size. The targets use a combination of high and low Emissivity patterns for checking detail acquisition by the system viewing the target. You can make one of these passive human targets very easily with nothing more than some wood board, silver (Aluminium) foil with adhesive backing (used in HVAC installations) and some Matt black enamel or acrylic paint. I will see if I can find the article on the target design and how to make your own version.
Fraser
-
This page provides images of various thermal imaging test charts used in front of a Blackbody IR emitter. Sadly no dimensions but they can possibly be worked out from the perimeter dimensions of the chart.
https://hgh-infrared.com/blackbody-targets/ (https://hgh-infrared.com/blackbody-targets/)
There is a Target calculator here….
https://www.ci-systems.com/calculator (https://www.ci-systems.com/calculator)
Fraser
-
CI Systems have a brochure on thermal test charts here……
https://www.ci-systems.com/Files/InfraRed%20Targets_May_2023.pdf (https://www.ci-systems.com/Files/InfraRed%20Targets_May_2023.pdf)
-
Inframet document on test targets that may be of interest…..
https://www.inframet.com/Data_sheets/Targets_IR.pdf (https://www.inframet.com/Data_sheets/Targets_IR.pdf)
-
The NPSA THERMAKIN human size test target that I mentioned. It is easy to DIY as detailed in the document.
https://www.npsa.gov.uk/resources/thermal-imaging-test-target-thermakin-manufacture-and-test-standard (https://www.npsa.gov.uk/resources/thermal-imaging-test-target-thermakin-manufacture-and-test-standard)
I attach the PDF.
-
the NFPA tests use a thermal version of the ISO12233 chart
https://www.graphics.cornell.edu/~westin/misc/ISO_12233-reschart.pdf (https://www.graphics.cornell.edu/~westin/misc/ISO_12233-reschart.pdf)
They only use the curved converging lines as seen top left, and then have a Matlab calculation operated on a series of still images taken of the LCD.
The difficult bit (assuming you want a correct rather than goodhigh answer) is making sure you do not keep counting 'resolution' in the aliased areas.
Simpler to make is simple converging lines of a fan chart (below) made by linking coarse and fine pitches to suit.
An old 'big bar / 4 bar' chart can be used presenting at pixel pitch. You get the full signal off the big bar and read off the signal of the 4 bars. You cannot cheat as the best result is lining up the chart with the pixels.
The 4 bar is the classic 'tank identification' test pattern used with presentation optics to set apparent range / size
-
Thanks for your inputs, an interesting read!
-
If you want to be super cheap, you can just use a bright halogen spot lamp or two pointed at a printout of a standard test chart. I've done this before using a test chart printed on several sheets of A1 paper for testing wide angle and lenses and lenses with a long fixed focus..
The uneven illumination doesn't make is super scientific, but it can be a good aid for focusing.