Whenever you take a photo, something derange the image: cars, bikes, other people and so on.
Like this:
But not every place is covered all the time !
So let's take 15 pictures of the same place (without moving the camera!):
Now we got uncovered pictures from every place but not at the same time.
All we have to do is to reassemble the fragments:
=>
I know, the image looks not so good, but it's just the prototype. It has only to show you the general idea.
This means in practice:
Take and store several pictures (maybe add RAM)
Add a microprocessor (simply a number cruncher)
and last but not least a self-timer (to avoid jiggles)
For the final picture we recalculate every pixel.
From each pixel we extract the color value:
1, 1, 2, 1,
1, 10, 1, 1,
1, 1, 3, 1,
4, 1, 1
The summation is 30.
The average is 2 (30/15), but as you can see in the table, the right value is "1".
(The example was made with particularly addition.)
= The clue is the algorithm to catch the most occuring number.
=>