Thermal image side by side with radiometric image?

[Abbott242]

Hi all,

I'm trying to interface with a new core via UVC, but the results I'm getting... Well, speak for themselves. I assume this is a result of the core sending radiometric and video data at the same time? This was just via using the default Windows camera app for a quick test.

I would appreciate any help in trying to diagnose the problem, as this is a first for me. Thank you!

-Abbott242

[Bud]

Try reversing the byte order in the right half of the image.

[Abbott242]

I'm not entirely sure what you mean... This is just a uvc camera - I don't think I can change each half of the image separately? But then again I'm not terribly experienced with software.

[LesioQ]

Big endian, etc.

https://en.wikipedia.org/wiki/Endianness

[IR_Geek]

what size is the array supposed to be?   one picture you show is 3636x1496 and the other is 2720x1496  ;  what is the digital bit depth of the camera ... 8 bit, 12 bit, 14 bit?    You should be able to calculate your raw digital counts.   ignore the color as that is just a color map being applied.

if possible set the camera in RAW without a color map (or just grayscale) and look at the native bit depth.   0-255 ; 0-4095; 0-16383;  even better if the camera has a test pattern so you can verify output.

[IR_Geek]

Also what was mentioned above about byte order.   Another is where the 'data' is stored inside the stream.   for 8 bit it could be packed into a 16 bit stream.  Then it must be pulled apart and applied correctly.  Could still be byte swapped.   if it's 12 bit in a 16 bit then need to verify where the zero are located ... first 4 our last 4.   If the camera vendor is a real PITA then maybe 2 zeros in the front and 2 at the end.   

OR the other data location is where the temperature calibration is being stored if it doesn't have a header for each frame.

[figment]

What camera model are you working with? I have a USB thermal camera that gives the same mirrored double image.