reply to post by CHRLZ
You've got very good points here and I fully agree with you about the three points you raised up there, especially when you said:
You cannot determine distance by supposed haze, unless you know exactly how bright or what colour the object is, in the first place.
We don't know what the actual colour / brightness /reflectivity of the object is, so it's impossible to determine this way if any of it's
appearance is due to haze.
The only time any useful determination from something's appearance can be done is if it contains areas that are too "dark/contrasted" to be more
than a given distance, given the haze effect should reduce that darkness/contrast. In this case we have the opposite - the object is very light/low
contrast and we have a bright day with LOTS of ambient light so if the object is reflective/partially transparent and bright whitish blue, then it
could be anywhere from right next to the lens to far off in the distance.
Let's see with examples how it works.
In a RGB photo, the Red channel is the less sensitive one to atmospheric haze.
Short Wavelength (450 nm)
Long Wavelength (630 nm)
Short wavelength radiation is scattered more sharply than long wavelengths by small particles in the atmosphere. This phenomenon is known as
. When there is an increase in atmospheric particles, short wavelength
scattering causes the appearance of haze. Landscape photographers traditionally use yellow or red long-wavelength filters to reduce atmospheric haze
with panchromatic black-and-white films.
Let's look now closely how blue haze impact red and blue channel in some examples.
This photo, for example, exhibit blue haze at various range, depends of the distance of the mountains:
Here's now a vertical radiometric cross-section of these mountains:
Now, see how the haze have a clear effect on both red and blue channel:
1: this is the sky part, the more the cross-section is close to the mountains, the more there are in the red channel, simply because there is a much
longer path for the scattered red light to arrive to the observer (than if it comes from overhead, for example), through the lower atmosphere, which
is where most of the aerosols are concentrated. The scattering effect is then magnified, causing more red light to be scattered than other any
2: the far away high moutains are particularly affected by blue haze. But sometimes there might be other particles in the air that are much smaller.
Some mountainous and sea regions are famous for their blue haze. Aerosols of terpenes from the vegetation react with ozone in the atmosphere to form
small particles about 200 nm across, and these particles scatter the blue light. Those particles could be as well droplets from the ocean and scatter
blue light as well.
The more distance there is between the subject and the observer, the more the atmosphere is thick, and the more there are blue light scattered with
3: here, the curves slowly decreases as we are getting closer to the observer.
4: finally, for the closer mountains, the red channel value is around 30 and the blue around 70. Note the important slope each time we change the
distance of the mountains (red arrows at the left of "2", "3" and "4")
5:note the means for the blue and the red channel (respectively 79 and 41)
Now, interestingly, the guy who've done the photo above
to correct this blue haze using
the red channel correction under Photoshop:
(BTW, that's what I've done in my previous post here
So, what about our radiometric cross-section, now?
Note that the curves of the two channels have now a completely different look, with a mean for both significantly lower (46 and 5)
Anyway, is all the above could apply to our "ufo" photo?
Firstly, here's what this photo looks like under blue channel:
... and red channel:
...to be continued...