FLIR One Developer / Blog / Topics / Thermal noise / Thermal noise pixels make temperature measurement difficult

Thermal noise pixels make temperature measurement difficult

Home Forums FLIR Mobile SDK Development Android Development Thermal noise pixels make temperature measurement difficult


This topic contains 3 replies, has 3 voices, and was last updated by Profile photo of Gavin Gavin 4 months, 3 weeks ago.

Viewing 4 posts - 1 through 4 (of 4 total)

  • Author


  • #11942

    Profile photo of David


    To the FLIR Mobile SDK development team.

    When using the Flir Mobile SDK to process thermal image streaming, thermal noise pixels often occur, making it very difficult to make accurate temperature measurements.
    You can confirm this by running the FlirOneCamera sample app distributed with the SDK by FLIR.

    I am measuring the highest temperature point on the thermal image, but the noise pixel is measured as the highest temperature point, causing an error in the real measured value, and there is a problem that may give wrong information to the user.

    When I checked it in the official app distributed by FLIR, I was able to confirm that no noise pixels were generated. There seems to be a difference in the way the thermal image is processed between the SDK and the official app.

    The thermal noise pixel issue is a very important factor in the effectiveness of the SDK, so we hope that the development team can quickly check it and improve it if problems are found.
    If it’s not a problem with the SDK, I would be grateful if you could tell me how to do it to get rid of the noisy pixels.

    Below are some sample images for this issue.

    * Test environments
    – FLIR One Pro LT 3rd gen, USB-C
    – Samsung Galaxy A7(2017) with USB-C port
    – Android 9
    – Tested by FlirOneCamera sample app and My app on developing.
    – MeasurementCollections -> MeasurementSpot

    * video clip about the issue above

    * noise pixels on the thermal image


    Profile photo of Mark


    There may be something wrong with your camera. I have not had this issue. You can programatically remove outliers or write a loop to identify which pixels are doing it and skip them from the value-pulling process loop.

    What is your application? I used this method for my app.


    Profile photo of David


    After testing several FLIR cameras (1 FLIR One Pro and 3 FLIR One Pro LT), I was able to find thermal noise pixels in all of them. And I can’t see the noise pixels in the FLIR One app distributed by FLIR Systems, so it doesn’t seem to be my camera problem.
    I understand the way Mark said, but I think there is a problem that the video may look unnatural due to the frames omitted in terms of fps performance.
    I am guessing if the SDK is causing this issue in the process of converting the thermal data to Bitmap.
    Okay, I may be wrong. So I am waiting for answers from FLIR development teams.


    Profile photo of Gavin


    Hi David,

    This issue has been reported to our technical report team and I will respond here with whatever information I receive.

    Thanks so much,
    FLIR Community Support

Viewing 4 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic.