Spectral Resolution

The term spectral resolution refers to the number and width of a specific wavelength interval (i.e., band or channel) in the electromagnetic spectrum to which a remote sensing system is sensitive. Spectral resolution describes the ability of a sensor to define the fine wavelength intervals in which electromagnetic energy reflected or emitted from Earth's surface features is recorded. In broad terms, the narrower the wavelength interval for a particular band or channel, the finer the spectral resolution. The spectral resolution of a remote sensing sensor depends primarily on two factors. The first is the number of bands or channels used and the second is the width or wavelength interval assigned to each band or channel. Higher spectral resolution can be achieved from using a larger number of bands and by assigning a narrower width for each band; however, the spectral position of a band is also important for achieving high spectral resolution. Narrow bandwidth is generally obtained by dispersing the spectrum of incoming radiation on to an array of detectors using prisms or diffraction gratings.

The spectral resolution of a standard black-and-white film is fairly coarse as the spectral sensitivity of the film is restricted to the incident energy in wavelengths extending over the entire visible portion of the spectrum and the various wavelengths of the visible spectrum (i.e., ∼0.4–0.7 µm [micrometers]) are not individually distinguishable. That is, black-and-white film records only the overall energy in the entire visible portion of the spectrum. Color film, on the other hand, is sensitive to the incident energy within the blue, green, and red wavelengths of the visible spectrum and therefore has relatively higher spectral resolution. It can discriminate between features of various colors based on their reflectance in each of these distinct wavelength regions.

Most satellite-based as well as airborne remote sensing systems designed for observing Earth's surface features record energy response in several discrete spectral bands at various spectral resolutions (Figure 1). These systems are commonly referred to as multispectral sensors. Landsat's thematic mapper (TM), for example, records electromagnetic energy reflected by and emitted from Earth's surface features in seven discrete bands covering the visible through thermal-infrared portions of the spectrum. The Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), on the other hand, is designed to provide spectral resolution of Earth's surface features in 14 discrete bands. It has 11 bands to cover the middle-infrared through thermal-infrared portions of the spectrum, compared with only three bands available to cover the same regions on Landsat's TM.

Figure 1 A graph showing the spectral properties of different features in visible through middle-infrared bands of the Advanced Spaceborne Thermal Emission and Reflection Radiometer

Source: Author.

Figure 2 Multispectral sensor versus hyperspectral sensor

Source: Author.

New advanced hyperspectral sensors are designed to detect reflected or emitted energy in hundreds of narrowly defined (i.e., narrow bandwidth) contiguous spectral bands (Figure 2). The Airborne Visible Infrared Imaging Radiometer (AVIRIS), for example, is designed to record reflected energy in the visible through middle-infrared portions of the spectrum by scanning 224 bands across a very narrow bandwidth. The Hyperion sensor on NASA's EO-1 satellite is [Page 2678]capable of providing spectral resolution of Earth's surface properties in 220 narrow spectral bands. A greater number of spectral bands available on hyperspectral sensors provide more spectral information about Earth's surface properties and are therefore capable of extracting even very subtle spectral information that usually cannot be resolved by multispectral sensors. Through these hundreds of very narrow-wavelength bands, fine spectral discrimination between different features on Earth's surface can be accurately achieved.

...