Frame color is too bright using GetThermal on raspberry pi

[윤여경]

I've installed GetThermal from the source and the below image is the result. I'm using Lepton 3.5 with Pure Thermal 2 on raspberry pi 5 and the same issue happens while using uvc-radiometry.py example code. Anyone who faced same issue? 

[윤여경]

I found that the min value on rpi is too low(3728.0) while on pc is 29552.0. Anyone knows why this happen? I'm using included libuvc and i didn't change anything...

[윤여경]

Max value of both pi and pc is ok tho

[윤여경]

Still don't get it why this happens but is there any chance that low voltage causes this issue? I'm using macbook pro power supply and it seems it can't supply enough power for pi.

[Nav]

It seems the issue is with the RPi 5 only. It works well on a  RPi 4. I guess  there is some issue with the USB connection on the new RPi 5 that is managed by a different chip.

[윤여경]

Thank you for your response. I should ask to my boss if i can test it on pi 4.

[Nav]

I am not sure but try to connect the Purethermal ground to RPi 5 ground if there are some noise in data.

[윤여경]

Finally figured out the reason why this happens! There is small noises on the frame which contain 0 value so this is the reason why the minimum range of temperature is too low like -273.15. My workaround is simple just replacing that 0 values to normal value. For instance i'm targeting the minimum temperature is 20 so replacing 0 to 29315. You can use this formula to convert Celsius to Kelvin `val` * 100 + 27315.

I'm attaching the complete workaround here. Hope it would help someone. I'm still don't know how that noises created.

[Griffin Covert]

I think those horizontal lines are the 0 value pixels throwing off the minimum range.

The solution to the little horizontal lines in the image is to run an EEPROM update on the Pi, then reboot:

sudo apt update
sudo apt full-upgrade
sudo rpi-eeprom-update -a

sudo reboot now