Some basics of image displays and color images

Learn Photo Editing

Learn Photo Editing

Get Instant Access

Most computer screens and image displays in use are 8-bit devices. This means that the displays can represent data projected on them with 28 = 256 different greyscale levels or data values of resolution. These greyscale levels can represent numeric values from 0 to 255 and it is common to only have about 200 levels actually available to the image display for representing data values with the remaining 50 or so values reserved for graphical overlays, annotation, etc. If displaying in color (actually pseudo-color), then one has available about 200 separate colors, each with a possible grey value of 0-255, or the famous "16 million possible colors" listed in many computer ads (see below).

On the display, the color black is represented by a value of zero (or in color by a value of zero for each of the three color guns, red (R), green (G), and blue (B)). White has R = G = B = 255, and various grey levels are produced by a combination of R = G = B = N, where N is a value from 0 to 255. Colors are made by having R = G = B or any combination thereof in which all three color guns are not operated at the same intensity. A pure color, say blue, is made with R = G = 0 and B = 255 and so on. You may have noticed that color printers have three (or four) colors of ink in them. They contain cyan, blue, and magenta (and black) inks, which are used in combination to form all the output colors. This difference (cyan etc. vs. RGB) in the choice of colors is simply because display screens mix light whereas printers mix ink to form specific colors.

Terms one hears but rarely uses in astronomy are hue, saturation, and brightness. Hue means the color of the image, saturation is the relative strength of a certain color (fully saturated = 1), and brightness is the total intensity of a color where black = 0. When you change colors (RGB) you are really changing the hue, saturation, and brightness of the image display. These three terms are fully explored in Gonzalez & Woods (1993) as well as in almost any text introducing image processing techniques.

Almost all CCD data obtained today have a dynamic range of much greater then 8 bits. Thus, in order to display the CCD image, some form of scaling must be performed to allow the image to be shown on a display with only 8 bits. A common technique (often performed by the software without user intervention) is called linear scaling. This type of scaling divides the entire true data range into say 200 equal bins, where each bin of data is represented by 1 of the 0-200 available greyscale levels. For example, if an image has real data values in the range from 0 to 100 000 ADUs, linear scaling will place the real data values between 0 and 500 ADU into the first scaled bin and will display them as a 0 on the screen. If your image is such that all the interesting astronomical information has real values of 0 to 2000, this linear scaling scheme will represent all the real image information for the values of 0 to 2000 within only 4 display bins, those having values of 0-4.

To avoid such poor scaling and loss of visual information, two alternatives generally exist: one uses a linear scaling but within a specific data window and one uses a different type of scaling altogether. The first option is accomplished by having the software again perform a linear scaling but this time using its 200 output display levels to scale image data values only within the data window of say 0 to 2000. Different scaling options allow for nonlinear modes such as log scaling, exponential scaling, histogram equalization, and many others. These are easily explored in any of the numerous image processing software packages used today.

A method commonly used to aid the eye when viewing a displayed image is that of interactive greyscale manipulation. You probably know this as changing the image stretch or contrast and perform it via movement of a mouse or trackball while displaying an image. The actual change that is occurring is a modification of the relation between the input data values (the 8-bits loaded into memory as chosen for display) and those output to the display screen. The software mechanism that controls this is called a look-up table or LUT. Some sample LUTs are shown in Figure C.1, where we see the relation of input to output data value. For greyscale images, all three color LUTs are moved in parallel, while in pseudo-color mode (see below), each color LUT can be individually controlled. Changes in the slope and intercept of the LUT control changes in the image brightness, contrast, and color. Terms such as "linear stretch" simply refer to a LUT using a linear transformation function.

Further information on general image processing techniques can be found in Gonzalez & Woods (1993), while additional details on image displays can

CONTRAST INCREASE

D CL

CONTRAST INCREASE

255

D CL

INPUT-»-SHIFT TOWARD HIGH GL

SLOPE EQUALS 1

INPUT

CONTRAST DECREASE

D CL

CONTRAST DECREASE

255

INPUT-»-SHIFT TOWARD LOW GL

D CL

INPUT-»-SHIFT TOWARD LOW GL

INPUT

INPUT

Fig. C.1. The top panel shows an example LUT, which converts input pixel values or grey levels to output grey levels on the video display. In this particular case, an input pixel value of 64 will be displayed with a grey level value of 80. Note that the relationship between input and output is not a single-valued linear transformation. The bottom panel shows various linear stretch operations. If the mapping passes through the origin and has a slope >1, the effect will be to increase the output contrast. A slope of <1 decreases the contrast. If the linear mapping intersects the y axis (i.e., the output value axis) the input values will be systematically shifted to higher output grey levels. Intersecting the horizontal axis causes a net shift to lower output grey levels. From Gonzalez & Woods (1993).

be found in Hanisch (1992). User manuals for astronomical image processing software packages such as IRAF and MIDAS often have numerous examples and routines useful for dealing with CCD imagery.

Was this article helpful?

0 0
Understanding Adobe Photoshop Features You Will Use

Understanding Adobe Photoshop Features You Will Use

Adobe Photoshop can be a complex tool only because you can do so much with it, however for in this video series, we're going to keep it as simple as possible. In fact, in this video you'll see an overview of the few tools and Adobe Photoshop features we will use. When you see this video, you'll see how you can do so much with so few features, but you'll learn how to use them in depth in the future videos.

Get My Free Video


Post a comment