Last night, I suddenly jumped on his head the idea of \u200b\u200bexplaining the operation of a digital camera (or at least the basics for beginners) that is used especially in this holiday season. Actually I can not even explain why I've come up with this idea, but so ...
A digital works basically the same way as an analog camera, with the main difference lies in the recording medium of photography itself: the film for the analog and the digital file in binary.
This difference obviously affects the construction of various devices within the machine.
As in the conventional machine, the camera records the image through the lens that focuses on a plan, called the focal plane . The focal plane is a film for traditional camera, while the digital is a grid of small boxes, called photodetectors, which "count" the number of photons, the particles that make up light, affecting the receiver . The greater the number of photons from the particular photodetector, the greater the electrical current generated. This current is then measured (the current is one of the few physical quantities that we know measure with high precision), thus obtaining a series of numbers, one for each grid point of the plan. At this point we know, in a way, how much light came in every point, enabling us to build an image that will be more faithful to the 'original', there will be many photodetectors in the grid.
However, as will have the most insightful, intuitive, with the information we have we can only reconstruct an image in black and white (or, rather, in "Grayscale"). In fact, following the procedures described, we can not distinguish the "color" of the individual photons that would enable us to reconstruct a full color image.
The problem is the fact that detectors are not able to distinguish colors. Or take all the photons without distinction of color or those who take a particular color (previously decided by the manufacturer). In both cases, however, we can not get the information needed to reconstruct the color image we want.
The first idea to make a color photo has been proposed by Maxwell (the same as the theory of ' electromagnetism). He made three identical black and white photographs in the three cases using three different filters: one red, one blue and one green.
overlay of three pictures (and then mixing various colors) Maxwell realized then the first color photos. Clearly, though, the inconvenience of the construction of a color photo.
Theoretically this process could be done mechanically: a series of three shots, where you change the filter every time. Basically the "change the filter every time" is the main problem. You can not instantly change the characteristics of individual photodetectors or apply a filter to quickly and mechanics. So what?
As you guessed, the system used in modern digital cameras use a filtering system similar to that proposed by Maxwell. There are actually several ways to achieve it, but the most common uses a so-called array of color filters, a pattern of filters (called Bayer ) shown:
With these filters we can produce distinct images, and incomplete, for the colors red, green and blue. The smartest have noticed that there is a predominance of green than the other two colors, and this follows in some way the sensitivity of the human eye: in fact our eyes are more sensitive to green than other colors.
We have almost finished our job. For now we only have 3 files incomplete for 3 colors. But we want a complete picture. How do we know how to complete the grid?
We pull to guess!
may seem a joke, but it works more or less like this.
Take, for example, only the green grid. We want to calculate how much green there is a pixel not covered by the grid. To do that we use a method called interpolation : exploit, ie, the information coming from the pixels adjacent to what we want to calculate. Then perform a sort of average intensity of light from the pixels that are re-washed in immediate vicinity of the particular pixel under study ... et voila, we have created a complete image for each color.
Obviously the system is considerably more complex interpolation, but it works more or less on the principle (called by some " demosaicing ") that I wrote.
We now show two examples of this process:
-> 
Vie seemed all that complex system? And if I told you that the sight of the human being works in exactly the same way (including the system of interpolation)?
This was more or less what I was going to say on the subject. I avoided going down on details, is not to weigh the subject, both for my ignorance on the details.
0 comments:
Post a Comment