10.5 How Do You Display a Digital Remotely Sensed Image?

Once the digital data is collected, the next step is to transform it into something that can be viewed on a computer screen, especially since many of these bands of energy are only visible to the sensor and not to the human eye. Sensors can only “see” a certain area of the ground at a time and will record measurements within that particular area. The size of this area on the ground is the sensor’s spatial resolution: it’s the smallest unit of area the sensor can collect information about. If a sensor’s spatial resolution is 30 meters, each measurement will consist of a 30 meter × 30 meter area of ground, and the sensor will measure thousands of these 30 meter × 30 meter sections at once. No matter how many sections are being measured, the smallest area the sensor has information for is a block that is 30 meters × 30 meters. Keep in mind, the spatial resolution of the image is going to limit the kind of visual information you can obtain from it. In an image with 4-meter resolution, you may be able to discern that an object is a large vehicle, but you won’t be able to go into any details about it. An image with 0.5-meter resolution, however, will contain much more detailed information about the vehicle.

We’ll discuss more about spatial resolution in Chapter 11, but for now, each one of these blocks of the ground being sensed represents the size of the area on the ground having its energy reflection being measured by the sensor. Owing to the effects of the atmosphere, there’s not a straight one-to-one relationship between the reflectance of energy and the radiance being measured at the satellite’s sensor. The energy measurements at the sensor are converted to a series of pixels, with each pixel receiving a brightness value, or BV (also referred to as a Digital Number, or DN), for the amount of radiance being measured at the sensor for that section of the ground. This measurement is made for each wavelength band being sensed.

Brightness values are scaled to fit a range specified by the sensor. We’ll discuss these ranges in more detail in Chapter 11 (with regard to a sensor’s radiometric resolution), but for now, the scale of values in a range represents the energy being recorded—from a low number (indicating very little radiance) to a high number (indicating a lot of radiance)—and measured as a number of bits. For instance, an 8-bit sensor will scale measurement values between 0 (the lowest value) and 255 (the highest value). These values represent the energy being recorded in a range from 0 to 255 (these values are related to the amount of energy being measured). For instance, all values in a band of an image could possibly only be in a range such as 50 to 150, with nothing reaching the maximum value of 255.

Each wavelength band that is being sensed is assigned a BV in this range for each pixel in the image. Brightness values (BVs) can be translated to a grayscale to be viewed on a computer screen. With the range of values associated with 8-bit imagery, values of 0 represent the color black and values of 255 represent the color white. All other integer values between 0 and 255 are displayed in shades of gray, with lower numbers being darker shades and higher numbers being lighter shades. In this way, any of the wavelength band measurements (including the wavelengths normally invisible to the human eye) can be displayed in grayscale on a computer screen (see Figure 10.8). For instance, a near-infrared image will have the brightness value for each pixel representing the near-infrared measurements at the sensor for that area.

If a sensor is measuring the visible portion of the spectrum and treating the entire 0.4- to 0.7-micrometer range as if it were one band, the result will be black-and-white panchromatic imagery. However, remote sensing devices are capable of sensing several wavelengths of the electromagnetic spectrum simultaneously. Some widely used remote sensing devices can sense several different bands at once—U.S. government satellite sensors are often capable of sensing 7, 10, or even 36 bands at the same time. Imagery created by sensing multiple bands together is referred to as multispectral imagery. Some technology can simultaneously sense over 200 bands of the electromagnetic spectrum, producing hyperspectral imagery.

When multiple bands of imagery are available, they can be combined so that the image may be viewed in color, rather than grayscale. Displaying imagery in color requires three bands to be displayed simultaneously, each shown in a separate color other than gray. This is simplifying the technical aspects a bit, but a display monitor (like a computer screen) is equipped with devices that can draw a pixel on the screen in varying degrees of red, green, and blue. These are often referred to in remote sensing as color guns (a name that refers back to equipment in old CRT monitors and tube televisions for displaying colors in red, green, and blue). If an 8-bit image is drawn using the red color gun, the values of 0 to 255 will not be in a range of black, gray, and white, but instead will be in a range of the darkest red, medium red, and very bright red. The same goes for the blue and green color guns—each gun is able to display an image in its respective color on a range of 0 to 255.

A color composite can be generated by displaying the brightness values of one band of imagery in the red gun, a second band of imagery in the green gun, and a third band of imagery in the blue gun. Each pixel will then be drawn using brightness values as the intensity of the displayed color. For instance, a pixel can be displayed using a value of 220 in the red gun, 128 in the green gun, and 75 in the blue gun, and the resulting color shown on the screen will be a combination of these three color intensities. By using the three primary colors of red, green, and blue, and combining them to varying degrees using the range of values from 0 to 255, many other colors can be created (using the 256 numbers in the 0–255 range, with three colors, there are 16,777,216 possible color combinations) to be displayed on the screen (see Figure 10.9 for examples of color formation).

To create colors, different values can be assigned to the red, green, and blue guns. Table 10.1 shows some examples of ways to combine values together to form other colors (keeping in mind that there are millions of possible combinations of these values). See also Hands-on Application 10.3: Color Tools Online: Color Mixing for an online tool that can be used to create colors from various 0 to 255 values.

When an image is examined, the color being displayed for each pixel corresponds to particular values of red, green, and blue being shown on the screen. For a remotely sensed image, those numbers are the brightness values of the band being displayed with that color gun. Any of the bands of an image may be displayed with any of the guns, but as there are only three color guns, only three bands can be displayed as a color composite at any time (Figure 10.10).

For example, the near-infrared band values could be displayed using the red color gun, the green band values could be displayed using the green color gun, and the red band values could be displayed using the blue color gun. Using this arrangement, a brown pixel can be seen on the screen. On closer examination, it’s discovered that the red gun (which has the near-infrared band) has a value of 150, the green gun (which has the green band) has a value of 80, and the blue gun (which has the red band) has a value of 20. These values correspond to the brightness values of their respective bands and combine together to form that pixel color. If you change the arrangement of bands and guns, you’ll see different colors displayed on the screen, but the actual brightness values will not change. Table 10.2 shows how the display color will change depending on which band is being displayed with which gun.

Note that while the pixel being displayed on the screen will change color depending on what band is being displayed with a different gun, none of the brightness values of that pixel will be altered. Different band combinations can reveal different information about objects in an image due to the varying brightness values of the bands for those items.

Several different types of composites can be formed using the various bands and guns. Whenever a composite has the red band displayed in the red gun, the green band displayed in the green gun, and the blue band displayed in the blue gun (the way natural colors would be shown), a true color composite is formed. A true color composite appears as you would expect a remotely sensed image to appear if you were looking down at the ground from an airplane (that is, the trees would be green and the water would be blue). Whenever the distribution of bands to guns deviates from this arrangement (some other combination is used besides the true color one), a false color composite is generated. For instance, all four of the combinations described in Table 10.2 would result in false color composites as none of them display the red band in the red gun, the green band in the green gun, and the blue band in the blue gun. See Figure 10.11 for a comparison of true color and false color composites.

With all of these different combinations possible, there is a standard false color composite that is used in many remote sensing studies. In this standard false color arrangement, the near-infrared band is displayed in the red gun, the red band is displayed in the green gun, and the green band is displayed in the blue gun. The false color composite image is similar to the CIR photos discussed back in Chapter 9. Objects that have a strong reflection of near-infrared energy (like healthy plants, grass, or other green vegetation) appear in a variety of bright reds, while water (which has very little reflection of the near-infrared, red, and green energy) appears as black. With composites, multispectral imagery can be displayed in color, regardless of what bands are being used, and these composites can be used for many different types of image analysis (see Hands-on Application 10.4: Comparing True Color and False Color Composites for an online method of comparing how objects on the ground are shown in true color and false color composites).