From: blockedofcourse@foo.invalid   
      
   On 11/20/2025 8:55 AM, Martin Brown wrote:   
   > On 19/11/2025 19:04, Don Y wrote:   
   >> On 11/19/2025 3:33 AM, Liz Tuddenham wrote:   
   >   
   >>> I think you might have to measure daylight and darkness observations   
   >>> differently.  The apparent temperature of the clouds will increase when   
   >>> they are in sunlight - but so will the apparent temperature of the space   
   >>> in between them, which is filled with illuminated dust particles.   
   >>   
   >> But, during daylight, I can distinguish between blue skies and white/grey   
   >> clouds.  I don't have to use the same wavelengths for all my observations.   
   >   
   > Although you can probably get away with it orange-deep red.   
   >   
   > A combination of a hot mirror filter (against hot IR) and a low pass wratten   
   29   
   > or 25 filter will make blue skies pretty much black and still leave clouds   
   > looking bright red. Contrast helps AI interpretation.   
      
   Remember, I can look at the sky every day and "remember" what I see.   
   This is important to allow the borders of the sky to be delineated   
   so earth-bound objects aren't interpreted in that field.   
      
   > Monochrome cameras tend to be cheaper than full colour. Beware that not all   
   > colour cameras react well to aggressive optical filters in front of them -   
   > Bayer demosaicing can go crazy faced with such weird data.   
      
   I'm going to start by playing with an outdoor COTS PTZ camera and see   
   what I can do to constrain the optics to a smaller portion of the field.   
   Again, as a human isn't viewing the signal, all I need to do is ensure   
   it is repeatable and "makes sense" to an image processing algorithm.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)