Ground Resolution

Let us recall that the linear ground resolution is derived from the spatial resolution, in radians, and the altitude. In the thermal infrared the spatial resolution depends upon the angle of view. For an angle of 1°, the spatial resolution is 17.45 mrad for 0.3°, 

Detectors that produce images, as a function of frequency 

Far and middle infrared I mm-3 um 3 X 10 -I0 Detectors with sweep system. Various types of image formation on CRT (cathodic tubes) 

Near infrared 0.7-3 nm 10 to3.85x 10 Photo, emulsions up to 1 nm and IR detectors with sweep systems. Excellent results with CRT. 

Visible 400-700 mu 7.5x 10 to3.85x I0 Films. Detectors with sweep system (multispectral), television. 

---

Figure 7. Ground resolution cell. Grazing angle is defined as /- a.

These observations must be kept in mind when passive microwaves are considered. For the earth sciences, the ground resolution is of vital importance. However, it is the author s feeling that one should seek neither too big nor too small a linear ground resolution. Ten kilometers is the extreme limit, but 1 km should also be the lower limit better still, between 500 m and 1000 m. Resolution finer than 500 m is the province of the aircraft, at lower cost. 

A number of Soviet satellites could certainly meet our needs despite very inadequate ground resolution unfortunately, it seems that the information gathered by these meteorological satellites does not belong to the public domain, as is the case for Nimbus, ESSA, TIROS, and ATS. ... 

Ten TIROS satellites were launched the first in 1960, and the tenth in 1965. With the exception of the last two, whose orbits were sub-polar, they were inclined 48 or 58° to the equatorial plane. All except four operate in the visible (TV) range the exceptions are TIROS 2, 3,4, and 7 which also work in infrared, but with very poor ground resolution. Though extremely useful for work in meteorological sciences, these satellites are of no interest to us. 

Earth scientists have little interest in the images received from the ATS. However, the ATS of the future will carry infrared radiometers whose space resolution of 0.1 milliradian will permit ground resolution of 3 to 4 km (altitude 30000 to 40000 km). 

When we achieve linear ground resolution equal to that of the infrared radiometers the passive hyperfrequencies will occupy the first rank. 

Two detectors receive, one in the 8.4-9.4 pm band, the other in the 10.2-12.4/im band, avoiding the O3 absorption region. The instant field of view is 0.6 mrad giving ground resolution of 600 to 630 m. The area scanned is limited to 400 km on each side of the field directly below the satellite it thus takes three to four days to cover the area scanned by the old Nimbus (HRIR) in a single orbit. 

As I see it, ground resolution is generally sufficient. Maps on the scale of 1 100000 can be obtained, allowing us to tackle our problems with unprecedented ease. However, the most important innovation is not that of resolution, but involves the placement of the two twin channels on either side of the dip (Reststrahlen) which has been discussed previously. In practice, the two detectors supply two radiance temperatures at the same instant for the same object. The thermal difference between these two channels, resulting from a difference in emissivity, enables alkaline rocks to be distinguished from acid rocks. If no difference is recorded (plant cover, snow, oceans) we get, as in the past, a radiance temperature distribution map. 

Spectral bands 1 to 4 benefit from ground resolution of 100 m and channel no. 5, a little over 200 m. It is hardly necessary to stress the importance of this multispectral spectrometer. However, I would like to draw attention to channel 5 (ERTS B, 1975 ) which transmits a thermal image of the regions overflown both day and night in an east-west band 180 km wide. The comparative study of the five spectral bands (four of them measure daytime reflectivity and the fifth, emitted infrared) permits identification of most terrestrial formations on a scale that allows detailed studies to be made. This leads to results that are not only technically satisfactory, but are comparable to those obtained by traditional methods. 

The TV camera of the I DCS system (Nimbus 3 and 4) has a wide-angle lens (103°) the focal length is 5.7 mm. On the ground, resolution is 3.7 km. Each image is composed of 800 lines and scanning is done in 3 20" or one-quarter second per line, of which 0.225 s is for the active phase (signal reception) and 0.025 s for the passive phase. 

We will take the case of THIR on Nimbus 4. You will remember that the radiometer is operating in two spectral bands, 10.5-12.5 /im (atmospheric window) and 6.7 /xm (water vapor band) (see Plate D, photo 1, opposite page 81). The latter channel measures the humidity of the upper atmosphere and stratosphere through it the jet stream can be detected, and displacements of the front systems tracked. The ground resolution of this channel is 27 km. 

Florida seen by Nimbus 5, SCMR nighttime thermal infrared, very high ground resolution (around 600 m), on 24 December 1972 (Orbit 173). 

Translated into ground resolution, this simple error in determining one of the two horizons causes a displacement of 5.4 km vertically below the satellite, and much more on the edges. 

Since ground resolution is about 8 km, the largest scale is that of 1 1000000, and even this involves some mental gymnastics. On the NMRT, the signals are separated. However, there is overlapping not only from orbit to orbit but also from line to line. 

Using the same reasoning for the ERTS, and rounding out the ground resolution to 100 m (channel 4) we perceive that separation of all signals, noise included, would... 

And this is the whole point what can we expect from satellites, and what part do aircraft have to play This leads me back to the concept of ground resolution which is obviously finer from an aircraft than from a satellite. I would thus like to examine, as soberly as possible, the relative superiority of these two carrying platforms, and answer the question in the heading. 

All maps but one illustrating this chapter were computer-produced on the original approximate scale of 1 1000000 which is the greatest possible having ground resolution of only about 8 km. 

Fig. 26. San Diego area (California) seen through a passive microwave radiometer. - Flight Convair 990, 10 May 1967 between 19 20 hrs (E) and 19 22 hrs (W), GMT. Linear ground resolution from 125 m (W) to 225 m (E). Passive hyper frequencies 19.35 GHz. Dotted area thermal anomalies.

With Nimbus 5 and ERTS, mapping to scales of 1 50000 and 1 25000 is perfectly possible. Multispectral photography and spectrometry open up new possibilities leading to results that are not only engineering achievements but may well have future applications for the disciplines concerned. In my opinion, development of a microwave radiometer must be considered merely a prelude to a later stage and not an end in itself, since ground resolution is always far too inadequate for earth sciences. 

This is exactly the satellite-airplane problem. I have said it before and I will say it again ground resolution of less than 100 m is not the province of the satellite. It costs too much. The airplane will do the work more cheaply and much better. The real future of remote detection, excepting the improvements mentioned above, will be a partnership between the satellite and the airplane, each one staying in its respective domain. What is more, the coming generation - the students - must be prepared for this real future. 

---