http://Science2.FarTooMuch.Info/digcam.htm is this page's URL.


Digital Cameras

Digital cameras work very much the same way as the eye. Both have a lens to focus an inverted image on a sensor array. Each sensor array consists of sensors that respond to different wavelengths. Both sensor arrays convert light intensity to a voltage that is transmitted to memory. In both memories the image is represented by intensities organized by location. Both compress the data for long term storage.

In the eye, the sensor array is called the retina. It consists of rods and cones. Most of the cones are in our central vision. There are three kinds of cones. Each type responds to different wavelengths of light. To accurately record a color image in a digital camera, there must be sensors that respond to the same wavelengths of light as the cones in our eye. Therefore for three types of cones, three primary colors are needed to represent, in combination, all the colors we see. All the colors we see are electromagnetic radiation with wavelengths between 400 and 700 nm. [A nm is a nanometer or billionth of a meter.] All these colors can be seen when a prism is used break up white light into its components. A rainbow also breaks up light into a spectrum.

Each of the three types of cones responds to a range of color. The rho cone responds to the red end of the color spectrum. It is stimulated by red light. The gamma cone responds to green light. The beta cone responds to blue light. Note that the rho cone's peak sensitivity is to yellow light, but yellow cannot be used as a primary color becuase the gamma cone also responds to yellow light. Therefore a red phospher is used in computer monitors to stimulate the rho cones. Also note that the rho cones of the eye also respond to the short wavelengths of violet light. It is for this reason we see violet in the rainbow.

Because most humans have three different types of cones, we can separately distinguish what we call the three primary colors of red, green and blue. A person [usually male] that has only 2 types of cones [or less] is said to be color blind. Most color blind males have a defective 'X' chromosone that gives them a yellow primary instead of both red and green. Note that the colors below are the primary colors of your color monitor which are the colors of its three phosphors that are as close to red, green and blue as the selection of phosphors will allow. Combinations of these primary colors give us yellow, cyan, magenta, white and all the other colors we see.

The closer the three color sensors of the digital camera match the spectral color sensitiviy of the eye, the more accurate will be the color image. For example, in most digital color cameras, the red sensors do not respond to wavelengths shorter than yellow. Therefore a digtal picture of a rainbow will have colors from red to blue. As a test, take a picture of the rainbow from the bottom side of a CD illuminated with a white light [halogen bulb is best] and see if violet exists. Of couse a violet dress that reflects red and blue light will appear as violet in both the eye and the digital image.

In the eye, most of the cones are rho & gamma. There are not as many beta cones. The gamma cones respond better then the rho cones. Most of the visual information of size and shape is determined by the gamma cones. Therefore in a color camera, half the sensors are the sensors that respond to green. The organization of sensors varies between cameras but the following represents a typical color camera grid:

Note that each green sensor represents one pixel. Around each green sensor are two red sensors and two blue sensors. Light at each pixel is represented the average of the two red sensors, the green sensor, and the average of the two blue sensors. Thus three bytes with a range of 0 to 255 are used to represent the intensity of red, green and blue for each pixel. In this example is an array of 32 x 18 to represent 576 pixels requiring 1728 bytes to store. Some professional cameras use more than 8 bits for each color in each pixel. Note that the ratio of 32 to 18 is the same ratio as high definition [HD] video. [either 1280 x 720 or 1920 x 1080]

How many pixels are enough? To answer that question I would like to look at the resolution of the human eye. When looking at a 5" x 7" photograph at a normal distance, it has the same field of view as 50 mm lens on a 35 mm camera. This then will be the reference field of view. At 20 feet it would be about 14.4' [4.4 m] by 9.6' [2.9 m]. It is also known from the Snellen Chart that at a distance of 20 feet, the average human vision can resolve letters made on a grid of squares about 0.07" [1.8 mm] on a side. The calculation of how many of these squares would occupy an area of 14.4' x 9.6' gives a total of about 4.1 MegaPixels. Half the population would need a higher resolution to not to see the visual quantising. It would take 20/20 vision to read capital letters printed with Bold Ariel font size 48 at 20 feet. A good digital camera with a 14.4' field of vision at 20 feet should be able to resolve these letters as well.

Blow up the image so the width of the big 'E' is 7" & use at 20'

Most digital cameras compress the data for storage using the algorith defined by the Joint Photographic Group. An average JPG file takes up much less space than a file of pixels stored as BMP or TIF. JPG breaks up the image in tiles and stores the average color of each tile, along with its shading from top to bottom and left to right. Higher harmonics of shading may also be stored. JPG is a lossy compression, meaning that there is a loss in picture resolution. A good quality JPG will be about 10% of the size of the original image. JPGs smaller then 3% usually show a significant loss of image resolution.

Image files can be transferred to computer either by removing the SD memory chip and reading it from an SD reader connected to the computer, or by connecting a USB [Universal Serial Bus] cable between the computer and the camera. The computer then sees the camera as another "disk" from which files can be read.

Note that mothers of color blind sons also have the defective 'X' chromosone. But they almost always have a 'good' 'X' chromosone as well. Those who have 4 different kinds of cones are called quadchromatic because they have 4 primary colors. So if a husband says two things are the same color and his wife says "No, they are not.", now you know why. Also note that some birds are quadchromatic because they have a UVA cone. In nature, there is much useful information for the birds and the bees in the UVA spectrum that they can see.




NEXT PAGE

HOME PAGE




This Web Site was designed by Russ Lemon