Remote Sensing - Image Formation (Cont.)
 

The remaining image formation techniques all use some form of electronic scanning, instead of a "camera", to produce an image. The optical-mechanical scanner, for example, grew out of early developments by military intelligence in thermal mapping. Reduced to its most basic components, this consists of a rotating mirror, which focuses the energy from a small spot on the ground onto an electronic detector, which measures the amount of energy. As the mirror rotates, it sweeps a very narrow strip perpendicular to the flight line. Each such sweep is called a scan line. Thousands of these scan lines, side by side, make up an image of the ground. This image can be from visible light (like a camera), or from the thermal energy emission from the ground itself. Thus, the infrared line scanner (IRLS) actually produces a heat map of the ground as it flies over the terrain [See Image 11]. This can be very useful for evaluating ground conditions, finding water sources, locating heat leaks from underground steam lines, etc.

Image 11: Infrared nighttime line scan image of the large petroleum refinery at Baytown, Texas; note that light tones are warm, and dark tones are cool; very apparent are the canals and ponds with warm waste water, and also the oil spill beside the tanker at the left. (photo:Texas Instruments, Inc.)

A modification of the basic optical-mechanical scanner allows it to record the brightness of the terrain in several different "bands" or wavelengths, at the same time. This can be imagined as producing an image showing only the blue, green, or red energy of the scene and perhaps several different images in various parts of the infrared. The device which does this is becoming very common and is called a multispectral scanner. One example of this is the U.S. civilian remote sensing satellite called Landsat. It has been operating continuously since 1972, and now produces digital data using eight different spectral bands.

One very useful application of this data is to allow computer mapping of land cover types. The digital image processing software can easily turn the satellite image into a thematic map showing categories of land cover such as: urban land, agricultural land, forest land, water, bare soil, etc. This type of product is typically called a multispectral classification map. This has been done for many years in all parts of the world. It can be very useful to assess the ground conditions prior to sending demining teams into the field to undertake operations. The Indian Remote Sensing (IRS) Satellite system offers another alternative for medium resolution multispectral and panchromatic data. The French satellite, SPOT, uses a related system, called a pushbroom scanner, to produce similar kinds of data. A satellite imagery sample series for the U.S. National Mall in Washington, D.C. is shown below. This is intended to give a comparative view of the different systems [See Images 12-17].

Image 12: Landsat Thematic Mapper (TM) color infrared image with its 28-meter pixels; note the 170-meter high Washington Monument in the right center of the image (Eosat/Space Imaging, Inc.)

Image 13: SPOT 20-meter color infrared image (Spot Image Corp.)

Image 14 4: SPOT Panchromatic 10-meter image (Spot Image Corp.)

Image 15 : Indian Remote Sensing Satellite (IRS) 6-meter image (Space Imaging/Eosat Corp)

Image 16: Russian KVR-1000 2-meter resolution film camera
(Spin-2, Inc.)

Image 17: Space Imaging Ikonos 1-meter panchromatic image
(Space Imaging/Eosat Corp.)

Digital multispectral scanners are also becoming more common for aircraft operations. One specialized version of the multispectral scanner, called a hyperspectral scanner, allows automated, detailed identifications of ground cover types such as: tree species or specific crop varieties. This is possible because, the hyperspectral scanner can create hundreds of wavelength bands, instead of just a few, like the Landsat, or the French SPOT satellite systems. These narrow spectral bands allow the exact characteristics of the reflections from a certain kind of vegetation to be defined and then searched for automatically in the data set. These operations, however, are very demanding in terms of computing power and operator skill.


Satellites

Table 1: Existing and planned remote sensing satellite systems.

Satellite

Operator

Type

Resolution (m)

Revisit (days)

Landsat 5

Space Imaging

Multispectral

30

16

Landsat 7

US Government

Panchromatic
Multispectral

15
30

16

IRS

India

Panchromatic
Multispectral

6
23

5
24

SPOT

CNES/SPOT

Panchromatic
Multispectral

10
20

1-4

RADARSAT

Canada

Radar

8-100

3-35

ERS

European Space Agency

Radar

30-50

3-35

JERS

Japan

Radar

15

4-45

IKONOS

Space Imaging

Panchromatic
Multispectral

1
4

3-5

OrbView

Orbimage

Panchromatic
Multispectral

1
4

3

Quickbird

EarthWatch

Panchromatic
Multispectral

1
4

2-4

SPIN-2

Russia

Panchromatic
Panchromatic

10
2

8

The newest tools for civilian satellite imaging are quite revolutionary, and just emerging at the beginning of the 21st Century. These include the Space Imaging, Inc. system called Ikonos, the OrbView Satellites from Orbital Sciences Corp., the QuickBird satellite from Watch, Inc., just to mention a few [See Table 1]. These systems typically involve one-meter panchromatic imagery, and four-meter color imagery, which is better spatial resolution than the traditional remote sensing satellites such as Landsat and Spot. Many different products and services are arising out of these new sensors, but at this time the imagery is too expensive to be of major use in the demining community.

At the beginning of this remote sensing section, radar was briefly mentioned as being a good example of an active sensor. (It sends out its own energy.) Radar has a variety of special benefits for aerial imaging. Because it provides its own illumination (microwave pulses of very short duration),it is useable day or night. Radar also has other very useful benefits. Of all the sensors mentioned above, it has by far the best atmospheric penetration. Depending on the type, some radar can even penetrate clouds, rain and snow. Radar has been used to map tropical rainforest areas, because of its ability to penetrate much of the vegetation canopy, revealing the ground below. Current and historical civilian radar imagery is available from RadarSat International of Toronto, Canada, from the European Space Agency using its ERS system and from the U.S. Geological Survey

Commonly used Radar Bands:

  • Ka Band: 40,000-26,000 MHz (0.8-1.1 cm wavelength)
  • K Band: 26,500-18,500 MHz (1.1-1.7 cm)
  • X Band: 12,500-8,000 MHz (2.4-3.8 cm)
  • C Band: 8,000-4,000 MHz (3.8-7.5 cm)
  • L Band: 2,000-1,000 MHz (15.0-30.0 cm)
  • P Band: 1,000- 300 Mhz (30.0-100.0 cm) 

Radar is also very sensitive to slight differences in topography, making it a very good tool for creating digital elevation data sets of the ground. One stunning example of this was the recent Space Shuttle mission to map the topography of most of the land area of the Earth. This was the SRTM, or Shuttle Radar Topography Mission. Processing the data has begun, but it is expected to take a year or two. The data is anticipated to yield a much more detailed and consistent data model for the topography of the Earth, especially for the nations of the developing world. This improved earth science data will be particularly useful for humanitarian demining operations where topography is an especially important land characteristic.

The most important variety of radar for remote sensing of the Earth’s surface is usually referred to as synthetic aperture radar, or SAR [See Images 18 and 19]. However, the old-fashioned name, sidelooking airborne radar, or SLAR, is more instructive. As the name implies, the sensor is looking out to the side of the aircraft. It creates a wide swath out to one side or the other of the aircraft (or spacecraft). Even from a great flying height, the imagery can have very fine ground resolution. As previously mentioned, the sensor system can also highlight moving objects using the moving target indicator, or MTI display.

Image 18: Synthetic aperture radar image of Mohawk, Arizona; flight line is left and right across the top of the image; note the extreme radar shadows from the steep hills, and the detailed presentation of the drainage network of this desert area. (photo: Goodyear Aerospace, Inc.)

Image 19: SeaSat Radar view of Los Angeles/Long Beach Harbors, California; heavy concentrations of larger buildings produce the strong radar returns in some areas; water is generally an area of no return (i.e., black). (photo: Jet Propulsion Lab, Pasadena, California)

Click here to take an interactive aerial photography self test.

Table of Contents Resolution >>