EEVblog Electronics Community Forum
Products => Thermal Imaging => Topic started by: soyd on October 26, 2023, 07:33:47 pm
-
Hello,
i'm building a thermal camera from scratch with a new Lynred Atto sensor and there occured some problems in the calibration process.
NUC calibration worked just fine with 10°C and 60°C blackbody full field of view exposure. But when it comes to radiometric calibration up to 150°C partial field of view exposure in 2m distance i see the blackbody afterwards persistent in the image.
Can someone tell me if it is it necessary to expose the factory new sensor to "high" flux before calibration (burnin?)?
Best regards
soyd
-
Never heard of that being of any use.
How long does the remnant image stay there ?
If you are back to a blank scene and with high gain you are likely to see some remnant image. This is because you will have heated up the pixel structures (not just the sensing part) to some extent with the 150°C so you can get some after effects as things cool back down again. That is both 'live' effects and that effect stored by any FFC calibration.
Bill
-
The pattern is persistent since about 36 hours.
The sensor is set to lowest possible gain. I already calculated a calibration with a scene temperature range of 10-60°C. With this calibration this comes out to ~0,3°C higher temperature at the remnant over the hole temperature range. At least it still shows linear behaviour just with a little offset change.
It`s an uncooled shutterless approach, so there`s no live FFC. The correction routine was already tested with an PICO sensor, where radiometric calibration up to 150°C was done over full FOV of the sensor at only 25cm distance. Afterwards even looking for hours at a hot scene in 2m distance no such remnants were permanently visible.
Now i changed my teststand for radiometric calibration to 2m distance and only a part of the image is covered by the blackbody. The only difference i see is that i don`t exposure the whole sensor to high flux.
Some gray hairs more...
-
Sounds like a question you should be asking Lynred, no? I can't imagine this sensor is less than 10k euros.
Already happened, but not being a direct costumer of them, it`s hard to get answers...
-
Some pictures, sorry for the different temperature scales. It seems that the bolometers are slowly recovering.
Picture burnin, 2days ago
Picture burnin2, 48 hours later
XGA@60fps, no denoising, no stacking