7.   FUENTES DE EXTRACCIÓN DE INFORMACIÓN TIPO RASTER

7.3.           Imágenes de radar

Continuando con la información primaria para el desarrollo de un SIG, y dentro las fuentes de extracción de información tipo raster, se encuentra la información proveniente de radares, éstos, al igual que las fotografías aéreas y las imágenes de satélite, son sensores remotos aunque no tan ampliamente utilizados en la recolección de información fuente.

7.3.1.                    Que son

La palabra RADAR, es un acrónimo de Radio Detection and Ranging. La pequeña cantidad de energía microondas emitida por el sol y la tierra, hace que los sensores remotos activos existan. Un sensor de microondas activo, primero emite un impulso de energía en el rango de 1 a 30 cm. y luego mide la cantidad de energía reflejada desde la tierra, ese el principio de funcionamiento de los radares. Esto es, que utilizando transmisores de radio de alta frecuencia, el radar emite un haz de radiación electromagnética, y los objetos que se hallan en la trayectoria del haz reflejan, según su capacidad de reflectancia, las ondas, y el radar recibe la energía reflejada.

El radar, no sólo indica la presencia y distancia de un objeto remoto, también puede determinar su posición en el espacio, su tamaño, su forma, velocidad y dirección de desplazamiento y su textura.

Los inicios de la tecnologia de radar se iniciaron en los años 30 con el objetivo de detectar barcos y medir su proximidad. El primer radar operativo fue desarrollado por los Británicos durante la segunda guerra mundial y era del tipo PPI (Plan Position Indicator), una antena giratoria que emitía impulsos de microondas durante sus 360° de giro;  La energía reflectada recibida, era almacenada como puntos de luz en un tubo de rayos catódicos, así los objetos de alta reflectancia como nubes de tormenta, aviones y barcos eran ubicados en términos de su dirección y su distancia del instrumento PPI, infortunadamente, este sistema no tenía la suficiente resolución para ser usado en aplicaciones cartográficas.

El Side Looking Airborne Radar (SLAR), fue desarrollado después de la Segunda Guerra Mundial, y con fines exclusivamente militares, es un sistema remoto de gran resolución, capaz de producir imágenes tan detalladas como las fotos sobre grandes áreas. Existen actualmente dos formas del SLAR, el radar de apertura real y el radar de apertura sintética (SAR). En los años 50 después de la guerra y de que la tecnología fue desclasificada se comenzaron a utilizar realmente no solo en detección si no para la obtención de imágenes.  El primer proyecto que involucró la utilización del SLAR en mapeo civil, se realizó en 1967, y fue un análisis detallado de la provincia del Darién en Panamá, éste proyecto tuvo gran impacto en ese tiempo porque esa región jamás había podido ser fotografiada completamente debido a la persistencia de las concentraciones nubosas en el área. Este proyecto, abrió las puertas al uso de los sistemas de radar para mapeo en todo el mundo. Desde principios de los años 70's, programas de mapeo extensivo se han desarrollado para la explotación de minas y petróleo.

En 1971, un estudio de radar se utilizó para determinar los límites reales entre Venezuela y sus países vecinos, este proyecto también permitió un inventario sistemático de las fuentes hídricas venezolanas, así como, el descubrimiento de los nacimientos de sus principales ríos, que hasta ese momento no se habían podido fotografiar, igualmente mapas geológicos fueron conseguidos como producto de las imágenes de radar.

También en el año de 1971, se comenzó el proyecto “Radam” (Radar of the Amazon), un proyecto de reconocimiento de la zona amazónica y parte del noreste de Brasil. Este ha sido el proyecto de radar más grande que se halla llevado a cabo, para 1976, más de 160 mosaicos de radar cubrían un área de 8´500.000 kilómetros cuadrados, entre otros estudios, estos mosaicos sirvieron para análisis geológicos, inventarios arborícolas, ubicación de rutas de transporte y explotación mineral donde se descubrieron importantes yacimientos, también se hicieron seguimientos de la actividad volcánica y de los grandes ríos.

En estas regiones tan vastas y con gran porcentaje de nubes, las imágenes de radar son la principal fuente de información para inventarios mineros, forestales, de recursos hídricos, rutas de transporte y determinación de sitios para la agricultura.

Las imágenes de radar, también se han utilizado ampliamente para monitorear la superficie de los océanos, para determinar las condiciones de viento, oleaje, y donde lo hay, hielo. Incluso, las imágenes de radar pueden ser utilizadas para estudiar, hasta cierto punto y bajo ciertas condiciones, los fondos oceánicos.

Al igual que los satélites para teledetección, los radares ya fueron introducidos como sensores remotos espaciales; el lanzamiento del “Seasat” en 1978, fue el primer proyecto de colocación de radares en el espacio, luego el Shuttle Imaging Radar (SIR), y también, otros experimentos soviéticos en los 80's. En 1991, tres satélites de radar, fueron lanzados en un lapso de 11 meses, estos fueron el “Almaz-1” propiedad de la Unión Soviética, el “ERS-1” propiedad de la Agencia Espacial Europea y el “JERS-1” propiedad Japonesa. Actualmente también existe el “Radarsat” Canadiense.

Grandes avances en el conocimiento del planeta Venus, donde las condiciones atmosféricas son absolutamente inhóspitas, se consiguieron utilizando imágenes de radar tipo SAR, tomadas desde la nave espacial Magellan.

Las misiones SIR Shuttle Imaging Radar sobrevolaron todo el planeta en un par de semanas, obteniendo espectaculares imagenes de 40 metros de resolucion lo que muestra la versatilidad de esta tecnologia.

Montes Atlas - Marruecos.

Santa isabel, Islas Galapagos

Que se ve de raro en esta imagen.??

Imaging and Non-Imaging Systems

Radar systems may or may not be imaging systems. A well known example of a non-imaging system is a doppler radar system. These systems are used (for example) to measure vehicle speeds by measuring the frequency, or doppler shift between transmitted and return signals.

Plan position indicators (PPI) produce a type of image. These radars use a circular display screen to indicate objects surrounding the rotating antenna. They are commonly used for weather monitoring, air traffic control and navigational systems. The spatial resolution of the view map of the area under surveillance is too coarse to be used in remote sensing applications.

Radar imaging systems used in remote sensing applications consist of an air or space borne antenna. These antennas transmit and/or receive radar signals in order to produce imagery at a fine enough resolution for image interpreters to identify physical features on the Earth's surface.

Black bulge under fuselage (radome)covers the radar antenna.

Active and Passive Imaging Systems

 
 
Active and Passive Radar Imaging Systems

Imaging radar systems can be active or passive. Active radar systems transmit short bursts or 'pulses' of electromagnetic energy in the direction of interest and record the origin and strength of the backscatter received from objects within the system's field of view. Passive radar systems sense low level microwave radiation given off by all objects in the natural environment.

Radar imaging systems such as ERS (European Remote Sensing Satellite), JERS (Japan Earth Resources Satellite), and RADARSAT-1 are active systems. They both transmit and receive energy. Microwave scanning radiometers only receive microwave energy. The Japanese MOS (Marine Observation Satellite) and JERS satellite systems employ microwave scanning units. Since the source is a very low level of electromagnetic energy, this type of data is prone to noise.

One advantage of active radar sensing systems is that, since they provide their own source of energy, they can collect data at any time of the day or night. Passive sensors - optical, thermal, and microwave - rely on receiving the naturally emitted or the sun's reflected energy from the Earth's surface.

Radar and the Electromagnetic Spectrum

The electromagnetic spectrum

Radio waves are one component of the electromagnetic spectrum (EM). This spectrum describes wavelengths and frequencies of energy. Ultraviolet, visible light, x-rays, microwaves and thermal energy are other examples of electromagnetic energy. The visible part of the EM spectrum contains blue, green and red light with wavelengths in the range of 400 to 700 nanometres. Infrared energy ranges from 700 to more than 100,000 nanometres. Microwaves are much longer, ranging from 1 mm to 1 metre. The longest microwave is approximately 2,500,000 times as large as the smallest visible light wave. Radar band designations range from the Ka band occurring between 7.5 and 11.0 mm to the P band range between 30 and 100 cm.

Any system transmitting and/or receiving energy from the Earth's surface is affected by the atmosphere. Water vapor, dust, smoke, airborne pollutants and other small particles close in magnitude to visible and infrared (VIR) wavelengths cause interference in the path between target and sensor. VIR wavelengths can be dispersed or blocked before they reach the sensors. Since microwaves are longer, they are not as affected by these types of small particle matter. Therefore, sensors transmitting and/or receiving microwaves are able to 'see' through haze, cloud, light rain, snow, smoke, and pollution. As a result, radar images can yield valuable information that is not available in VIR images.

Atmospheric Penetration

Radar Systems in Remote Sensing

Radar Systems in Remote Sensing

In order to understand how a radar imaging system produces imagery, a comparison with optical imaging systems is useful. Photographs or scanned images are the product of systems which use visible light and near infrared radiation and are the result of near instantaneous exposure. In contrast, radar imagery is produced by recording microwave pulses travelling to and from a target area over a period of time.

The optics-equivalent in a radar imaging system is a long, rectangular antenna which transmits and receives microwave energy. Resolution, which is the ability of a system to differentiate between two closely spaced objects, is dependent on focal length in optical sensors and in the along-track direction antenna length in radar systems. Antennas are analogous to lens systems in that a long antenna can be compared to a telephoto (long focal length) lens, while a shorter antenna is similar to a wide angled (short focal length) lens. To continue the analogy, a long antenna provides a detailed, or high-resolution image of a small area, while a short antenna provides an image of a large area with less detail.

Range Resolution

Resolution in a radar system is controlled by the signal pulse length and the antenna beam width. The signal pulse length dictates resolution in the direction of energy propagation. This is referred to as the range direction. Shorter pulses result in a higher range resolution. The width of the antenna beam determines the resolution in the flight or azimuth direction. The beamwidth is directly proportional to radar wavelength and is inversely proportional to the length of the transmitting antenna. This means that resolution deteriorates with distance from the antenna. In order to have a high resolution in the azimuth direction the radar antenna must be very long.

Resolution in the flight or azimuth direction

Remote sensing radars can be divided into two categories - real aperture and synthetic aperture radars (SAR). Real aperture radars transmit and receive microwave signals with a fixed length antenna. They are limited in their ability to produce resolutions fine enough for most remote sensing applications, simply because it is difficult to transport a very long antenna. To solve this problem synthetic aperture radars (SAR) were developed. SARs have physically shorter antennas, which simulate or synthesize very long antennas. This is accomplished through modified data recording and signal processing techniques.

Imaging Geometry

Photographic system geometry

The geometry of radar imaging systems and photographic systems or VIR imaging scanners is very different. Where photographic systems and most scanners employ a central, downward looking sensor and symmetrical geometry, radar imaging systems are characterized by a side looking sensor and asymmetrical geometry.

Scanning system geometry

Radar system geometry

Incidence Angle

Incidence angle describes the relationship between radar illumination and the ground surface. Specifically, it is the angle between the radar beam and a target object. The incidence angle helps to determine the appearance of a target on an image.

A local incidence angle can be determined for any pixel on an image. Trees, rocks, buildings, other structures and different terrains create changes in the local incidence angle. This in turn causes variations in pixel brightness.

Local incidence angle influence on pixel brightness

Due to their greater altitude, satellite incidence angles vary less than airborne incidence angles. This leads to more uniform illumination on satellite images than airborne radar images.

Incidence angle variation from aircraft versus satellites

Slant Range and Ground Range

H_n = flying height    ß = depression angle
Slant Range and Ground Range

Slant Range Image


Ground Range Image

Radar imaging systems record the differences in travelling times between return signals. The distance between an object and the antenna is equal to the speed of propagation of the wavelength through the atmosphere multiplied by the time it takes to reach the antenna. This, of course, is the relationship between the speed of EM radiation, time taken and distance travelled. A direct relationship between the slant range and the ground range also exists. Since we know the angles at which the microwaves are propagated, we can use trigonometry to calculate the ground range.

In order to view a radar image in the more recognizable ground range configuration, a geometric correction between the two distances is made.

Geometric correction

R_o, ..., R_k, ..., R_n = slant range (km)
H_n = satellite altitude (km)
= illumination angle
= Earth centre angle

The slant range to ground range distortion is much more pronounced in airborne SAR systems than it is in satellite SAR systems. This is a result of the difference in depression angles and the range of the depression angles between airborne and spaceborne SARs. The slant range to ground range correction may not be necessary in order to create an effective stereo pair from satellite SAR imagery.

Radar Imagery Versus VIR Imagery

How does radar imagery differ from VIR imagery?

A quick glance comparing radar imagery with aerial photography or VIR satellite imagery will reveal obvious differences. Imaging geometry and electromagnetic wave properties together produce the very different appearances of a radar image, an aerial photograph or a VIR satellite image. The following sections demonstrate the effects of imaging geometry and microwave properties on a typical radar image.

VIR image (left) and radar image (right) comparing radar imagery with aerial photography or VIR satellite imagery

Relief Displacement

Vertical structures on radar images, and aerial photographs or VIR satellite images appear to be very different. The most obvious difference is that relief displacement is in opposite directions. On aerial photographs and other optical sensor images, relief displacement falls away from the nadir point because the top is imaged further from nadir than the base of a structure. In radar images the top of a structure may be imaged before the base. Thus, the relief displacement falls towards the nadir. Relief displacement will be greater in slant range than ground range due to the fact that the image is more compressed in a slant range presentation. Relief displacement is also most pronounced at near range.

Relief displacements occur in opposite directions for optical andSAR sensors

Range direction distortions on radar images are comparable to those encountered in oblique photographs. But, as can be seen from the previous figure, relief displacements occur in opposite directions for optical and SAR sensors.

The radar perspective represented on an image is portrayed as being orthogonal to the radar direction. Consequently, a viewing angle "" of less than 90°, usually employed in SAR systems, has approximately the same effect as an equivalent angle of "90°-"; for oblique optical viewing.

Relief displacement is in opposite directions

A = relief displacement in the direction away from the optical sensor

B = relief displacement toward the radar sensor

= incidence angle

Keep in mind this relationship. It will help you in understanding the impact of geometry on SAR imagery. For example, it is important when considering stereo configuration and the positioning of SAR stereo image pairs.

The four characteristics resulting from the geometric relationship between the sensor and the terrain that are unique to radar imagery are foreshortening, pseudo-shadowing, layover, and shadowing.

Foreshortening (A'B') is the effect by which the foreslopes of hills and mountains appear to be compressed. The image of foreslopes will therefore appear brighter than other features on the same image. The greatest amount of foreshortening occurs where the slope is perpendicular to the incoming radar beam. The base, slope and top of the mountain will be imaged at the same time and will be superimposed on the image. Foreshortening can be minimized by using a less sharp incidence angle. However, lower incidence angles allow for more shadowing to occur on the image.

Foreshortening
While the hill slopes AB and BC are equal, the foreslope (AB) is compressed (A'B') much more than the backslope (BC) is compressed (B'C'), due to the radar imaging geometry.

Pseudo-shadowing is an effect by which the backslope of hills and mountains appear to be expanded. It is the result of return signals spread out over a larger distance (A', B') than the actual horizontal distance (A, B). This dispersed return is not always detectable (Leonardo, 1983).

Pseudo-shadowing

A', B' = return signal spread    A, B = actual ground distance               A'< A   B' > B

Layover is the effect where the image of an object appears to lean toward the direction of the radar antenna. It is the result of the tops of objects or slopes being imaged before their bases. Layover effects are most severe on the near range side of images.

Layover
While the mountain slopes AB and BC are equal, the radar imaging geometry dictates that the radar-facing slope (AB) wil be imaged (B'A') as leaning toward the radar. This is due to the mountain top (B) having been imaged before the base(A) due to RA > RB.

Shadow is very useful to image interpreters interested in terrain relief. As discussed in the first chapter, shadowing is one of the psychological cues used for depth perception. Radar shadows produce a 3-D effect without the use of a stereoscope.

1 = Shadow area not imaged     2 = Radar shadow on image


Example of radar shadow effects under large incidence angle (>45°) illumination.

When terrain slopes are greater than the depression angle, true radar shadows mask down range features. In this case, slopes facing away from the radar antenna will return very weak signals if any. This results in dark or black areas on the image. In areas of high relief, as the depression angle becomes shallower, shadow length increases with range. The shallower the depression angle is on such terrain, the more information will be lost.

Microwave properties

Objects on the Earth's surface react differently with electromagnetic energy. The strength of reflected energy, which is recorded in order to produce a radar image, depends on factors such as:

It is important to note that microwave reflections from the Earth's surface are not related to their counterparts in the VIR section of the EM spectrum. Surfaces that return a strong signal and are bright in a radar image may return a weak signal in the VIR range of the spectrum and appear dark on a photograph, Landsat or SPOT image.

Surface Roughness

Surface roughness is determined with respect to radar wavelength and incidence angle. A surface will appear to be smooth if its height variations are smaller than one-eigth of the radar wavelength. In general, a rough surface is defined as having a height variation greater than half the radar wavelength. Surfaces will appear to have a greater or lesser degree of roughness, depending on which designated radar bandwidth is used for imaging. In terms of a single wavelength, a surface appears smoother as the incidence angle increases. On radar images, rough surfaces will appear brighter than smoother surfaces composed of the same material.

A = antenna; h = height variations of surface;  = radar wavelength.
Smooth surface; specular reflection; no return.

A = antenna; h = height variations of surface;  = radar wavelength.
Intermediate roughness; mixed scatter; moderate return.

A = antenna; h = height variations of surface;  = radar wavelength.
Rough surface; diffuse scatter; strong return.

Surface roughness influences the reflectivity of microwave energy. Horizontal smooth surfaces that reflect nearly all incidence energy away from the radar are called specular reflectors. These surfaces, such as calm water or paved roads appear dark on radar images.

Radar reflection from various surfaces

Rough surfaces scatter incident microwave energy in many directions. This is known as diffuse reflectance. Vegetated surfaces cause diffuse reflectance and result in a brighter tone on radar images.

Moisture Content

The complex dielectric constant describes the ability of materials to absorb, reflect and transmit microwave energy. The dielectric constant increases with the presence of moisture in a material.

Moisture content changes the electrical properties of a material, which in turn affects how the material will appear on a radar image. The reflectivity, and therefore image brightness of most natural vegetation and surfaces increases with greater moisture content. Consequently, soil moisture maps can be derived from radar backscatter.

Irrigation / soil moisture influences.

Test site at Outlook, Saskatchewan showing potato fields at pre-emergence stage;
C - VV airborne radar;
A = irrigated field,
B = non-irrigated field.

Soil moisture map
near Altona, Manitoba using C band radar data.

Microwaves may penetrate very dry materials such as desert sand. Both surface and subsurface properties affect the resulting scatter. In general, the longer the radar wavelength, the deeper it will penetrate dry material.

Fading and Speckle

Fading and speckle look like grains of salt and pepper randomly distributed on an image. Fading and speckle are noise-like processes inherent in coherent imaging systems. Radar imagery is created with coherent radar waves which causes random constructive and destructive interference producing random bright and dark spots in radar imagery. Fading and speckle can be reduced by averaging adjacent pixels on radar images, or by designing the antenna to use a multiple look imaging technique.

Coherent radar waves

Single look image - high spatial resolution and speckle


Twenty five look image (five looks in each of x and y directions) - spatial resolution and speckle have been reduced

Fading and speckle can be reduced by averaging adjacent pixels on radar images, or by designing the antenna/processor to use a multiple look imaging technique.

7.3.2.                    Ingreso de imágenes de radar al SIG

No puede olvidarse el motivo por el cual se están estudiando estos últimos temas; la razón es que todos son fuentes de extracción de información para un SIG.

El ingreso de información proveniente de sensores de radar, siempre y cuando no sea análoga, puede hacerse como la de la demás información digital tipo raster. El estudio y el análisis de imágenes de radar, son relativamente complejos, y no es muy común el software para el análisis de las mismas, sin embargo cuando pueden conseguirse y analizarse, las imágenes de radar ofrecen muchas ventajas, particularmente en zonas donde otros sistemas remotos no funcionan debido a la alta nubosidad, ya que el radar atraviesa sin ninguna dificultad los cúmulos nubosos.

El tipo de análisis que comúnmente se realiza es el mismo que se le haría a una imagen de satélite normal aunque en realidad se podrían cometer errores al hacerlo ya que las imagenes de radar tiene algunas características como las "sombras"