Remote Sensing and Image Processing - Essay Example

Remote Sensing and Image Analysis Remote sensing is the process of obtaining information about the earths surface from an overhead perspective (Campbell 2002). The process in Figure 1 starts with a signal from a source (A or D). The signal travels through the atmosphere (B) and becomes incident on a target object at the earths surface(C). Radiation reflected from, or emitted by the object again travels through the atmosphere (C). The signal reaches a sensor (D), a device that receives and records the signal. The resulting data is passed on to a satellite data relay station and then to a land-based station, or directly to a land-based station (E). Data is processed (F) before being provided to end users, who in turn, will conduct their own analysis for their specific purposes (G). Physical principles of electromagnetic radiation The signal commonly used in remote sensing is electromagnetic radiation. These waves have different wavelengths and frequencies, but travel at a uniform velocity. Waves with longer wavelengths have smaller frequencies, and vice versa. The different electromagnetic waves, arranged by increasing wavelengths, are the gamma rays, x-rays, ultraviolet rays, visible light, infrared, microwaves and radio waves. For remote sensing, the commonly used ranges of the electromagnetic spectrum are visible light, infrared, and microwave (Lillesand et al. 2008; Sabins 1987). Electromagnetic radiation may be reflected,refracted or absorbed (Campbell 2002; Lillesand et al. 2008). Reflection occurs as incident radiation bounces back from the surface of a material. On a flat surface, the angle of incidence is equal to the angle of reflection. However, a rough surface will cause incident radiation to be scattered or reflected in different directions. The velocity of a wave usually changes as it crosses the boundary between different materials. This change causes the wave to bend or refract (Campbell 2002). The rainbow pattern produced as white light passes through a prism is a result of refraction. Incident radiation can also be absorbed by, stored in, and later emitted by materials (Campbell 2002). In Figure 1, incident radiation comes from the sun (A) or an artificial source (D). Before it reaches a target object at the earths surface (C), it passes through the atmosphere (B). The air molecules in the atmosphere prevent particular wavelengths from reaching the surface through reflection, scattering or absorption, and allow other wavelengths to be transmitted to the surface. This screening effect is referred to as atmospheric spectral transmittance (Sabins 1987). Strongly transmitted wavelengths will high transmittance. For remote sensing of objects at the earths surface, wavelengths with very low atmospheric transmittance are not used (Sabins 1987). Materials at the earths surface (C) reflect incident radiations or emit their own radiations towards a sensor (D). These radiations will depend on the material properties such as color, temperature and surface smoothness. The differences in these radiative properties allow remote sensing operators to distinguish one material or surface type from another (Lillesand et al. 2008). Using the visible light spectrum(~0.4-0.7μm), a forest can be distinguished from an ice-covered terrain. Forests will mostly reflect visible light with wave length corresponding to the color green, absorb all others, and thus appear as green. In contrast, glaciers will reflect all visible wavelengths and thus appear as white (Sabins 1987). Surface temperatures ranging from ~220K to ~1000K can be distinguished by sensing thermal infrared radiation in the 3-5μm and 8-14μm range . Cooler surfaces register peak radiative intensities at larger wavelengths (Sabins 1987). Surface smoothness can be mapped using differences in scattering . Microwave signal incident at an angle of less than 90 degrees to a very smooth surface will be completely reflected, without any signal going back to the source along the original line of incidence. This returned signal is called backscatter (Campbell 2002). In contrast, a rough surface produces significant back scatter. Bodies of still water have zero backscatter. Where the angle of incidence is 90 degrees, the surface will have 100% backscatter. Satellite orbit selection Data generated from sensors will be influenced by the orbit of the satellites that carry these devices. Thus, the selection of satellite orbit will depend on the objectives of the remote sensing work (Campbell 2002). The orbit can be characterized in terms of its altitude, and its orientation and rotation with respect to the Earth (Campbell 2002; Sabins 1987). Altitude influences the coverage area as well as the resolution of images being produced. For a given sensor, higher altitudes will allow for a larger ground-projected instantaneous field of view (GIFOV) and thus allow the coverage of a larger area at a single time. However, the spatial resolution decreases as a result of zooming out. High-altitude platforms are well suited for remote sensing work which require the coverage of large areas at a time but do not require high spatial resolution. On the other hand, low-altitude platforms are better suited for work on smaller areas requiring high spatial resolution (Campbell 2002). A satellite may have an orbit parallel to or inclined relative to the equatorial plane. Orbits with high inclination are called polar or near-polar orbits. Sensors whose orbits are parallel to the equatorial plane have less coverage of higher latitudes than those with polar or near-polar orbits. Thus, remote sensing platforms which require a global coverage generally make use of polar or near-polar orbits(Campbell 2002). Satellites which have periods that match the rotation of the earth are called geosynchronous (Lillesand et al. 2008). A geosynchronous satellite, when viewed from the ground, will thus have a constant position in the sky at the same time of every day. A geostationary satellite (Lillesand et al. 2008) is a geosynchronous satellite in an equatorial orbit. This means the satellite remains stationary relative to the earth. This type of satellite is used when there is a need for continuous monitoring of a particular area, such as for weather forecasting. Satellites may also be sun-synchronous(Lillesand et al. 2008). These satellites pass a specific area at exactly the same time of the year. This type of orbital period is best for remote sensing work that requires several years of observations. It eliminates the need for corrections due to seasonal variations in illumination. Other inclined orbits periods shorter than the earths rotation. As a result, these satellites pass a particular area at different times of the day. Although illumination conditions for a specific area may vary, it has the advantage of getting an image of an area several times a day. At the same time, it has larger spatial coverage and greater resolution than those with geostationary orbits. This system is best for remote sensing work requiring global coverage and frequent, but not necessarily continuous observations (Campbell 2002). Sensor characteristics Sensors may be characterized as passive and active (Campbell 2002; Lillesand et al. 2008; Sabins 1987). Passive sensors detect signals that are either reflected by the target object from a natural source or naturally emitted by the target source. On the other hand, active sensors detect signals reflected by the target object from an artificial source. Cameras used for aerial photography are passive sensors because these devices make use of the natural illumination from the sun. Active microwave sensors generate their own microwaves which are directed at the target objects and then reflected back to the sensors. Spectral characteristics refer to the range of the electromagnetic spectrum which sensors are able to detect and the spectral resolution (Campbell 2002). Sensors with a narrow range usually have greater spectral resolution within the given range, while wider ranges offer lower resolution. Geometric characteristics determine how sensors receive the signals . These include the lens distortion, field of view (FOV) and instantaneous field of view (IFOV) (Campbell 2002). Pre-processing and analysis techniques (F and G in Figure 1) After the sensors have collected data from the earths surface, certain steps are undertaken to make the information useful to end-users (Sabins 1987) . One of these processes is image correction, which involves eliminating signal noise and geometric distortion. Georeferencing is carried out to accurately identify the location of the image( Sabins 1987). Analysis is facilitated by data visualization. The human eye can only perceive the visible part of the electromagnetic spectrum. Hence, infrared and microwave data need to be represented by an appropriate visible spectrum (Campbell 2002). For thermal remote sensing, higher temperatures are assigned to warm colors whereas lower temperatures are assigned to cool colors. Data from active microwave remote sensing is represented in gray scale. Zero backscatter is black, full backscatter is white, and in between these two are the various shades of gray (Campbell 2002; Lillesand et al. 2008; Sabins 1987). Interpretation of remotely-sensed images should be supported and validated by ground-truthed information. Ground-truthing may be done by actual fieldwork or by review of existing data and maps derived from independent methods(Campbell 2002; Lillesand et al. 2008; Sabins 1987). Using GIS, different remotely-sensed images of an area can be laid one on top of another to facilitate classifications based on the different parameters represented by the different images (Campbell 2002; Sabins 1987). Reference List Campbell, J.B.. 2002, Introduction to remote sensing, 3rd edn, Taylor & Francis. Lillesand, T.M., Kiefer, R.W. & Chipman, J.W. 2008, Remote sensing and image interpretation, 6th edn, John Wiley & Sons. Sabins, F.F. 1987, Remote sensing: principles and interpretation, 2nd edn, Freeman. Read More