Digital Pictures

To take a digital picture is more than just pointing and clicking. After you clicked the picture, the camera creates an image of what you took by assigning a value, in this case, a color, to a pixel. These pixels are organized in a grid in such a way that your mind combines them into a picture, and not a bunch of individual pixels. This process is called digitization, which is an application of the Riemann Sum Method.

To see how this works let’s take the image of this fish:

Fish Base.png

To you and me, we see a clown fish, but the computer is actually showing is just a series of pixels that are colored in such a way that your brain interprets it to be a clown fish. In fact, if you were to zoom in on the fish, you will begin to see those pixels.

To see how the process of digitalization works we’ll overlay a grid of squares over the fish as follows:

Fish Grid 1.png

Then each square is examined and assigned a color. In this case, if more than 50% of the square is a certain color then the square is assigned that color as its value. Which produces the following figure.

Fish Digitized 1.png

The above figure bears a slight resemblance to a fish, but for the most part, if I didn’t tell you that this was a fish then you probably would not be able to see it. To get a more accurate picture more, smaller pixels are required. In the next grid, the squares are half the size they are in the original, resulting in four times as many pixels as in the first grid.

Fish Grid 2.png

Applying that same method as above yields the following image.

Fish Digitized 2

The looks more like a fish, and we are beginning to get more details. The fish’s eye is present as well as its fins are beginning to be delineated. If we were to continue creating smaller and smaller pixels then we would get a more detailed fish. In fact, the fish’s image, which you looked at, contains approximately 66,000 pixels.

While 66,000 pixels may seem like a lot, modern computing power is measured in the teraFLOPS. A computer that can do only a single teraFLOP can compute 1,000,000,000,000 computations in a single second. So, it would take a computer 0.000000066 seconds to preform all of the calculations required to render the 66,000 pixels.