Foreshortening and Layover - Lecture Note - Completely Remote Sensing tutorial, GPS, and GIS - facegis.com
Foreshortening and Layover

The look direction, along which the radar beam is traveling, can always be determined by the position of bright slopes in hilly terrain - like those that are sunlit indicating where the Sun is azimuthally, the radar platform will be to the right (looking in the opposite direction) at approximately 90° to the pattern of elongated bright slopes.

Unlike solar illumination which, coming from a distant source, sends its signal as continuous parallel rays of light (photons) onto a sensed surface, radar sends a discontinuous (intermittent) series of photon pulses from a point source that then spreads out as an angular beam. Moreover, the return signal is closely tied to the transmitted signal in that the travel time to the target and back relates to the history of the pulse. The position of any part of the target relative to the near and far ranges affects its dimensions. As a generality, the width of a surface, be it horizontal or inclined, appears to increase as the beam spreads from near to far.

For topographic surfaces other than flat, the irregularies can have a significant effect on the appearance of the resulting image. In radar images where the terrain is quite mountainous with high relief, unusual geometric characteristics may appear. Hill or ridge slopes facing the radar are subject to a distorted appearance. The terms layover and foreshortening apply to this appearance. Both are expressed as a compression or "thinning" of slopes on the facing (bright-toned) side and an elongation on the side that is shadowed. In most instances, layover and foreshortening produce the same end result visually. This slope displacement is more pronounced in the near range part of a scene than in the far. This radar image is characteristic of this distortion effect:

Layover, and some foreshortening, causing topographic distortions, as seen in this radar image.

The image above is typical of the layover seen in aircraft radar images of mountainous terrain. The beam is directed from right to left. The visual pattern gives the impression that the mountains, whose opposing surfaces have similar angles, are analogous to the special topographic feature known as "flatiron" form (hogbacks along the Rocky Mountain Front resulting where rock units are dipping in one direction are this geomorphic type).

All foreslopes (those facing the incoming beam) are shortened to some extent in radar images. Visually, these slopes appear distorted, with the facing slopes seeming to lean toward the radar platform, as though they are steeper. As depression angles increase, the geometry makes the slope lengths (top to bottom) appear to progressively decrease, thus increasing the degree of foreshortening. Slopes on the opposite side of mountains with ridges will generally not be illuminated and are thus rendered as shadows (no signal returns and therefore dark). Radar-blocked shadows become wider and darken as look angles increase, so they too are distorted from their projected dimensions. These ideas are summarized in the following diagram, adapted from a figure by A.J. Lewis, 1976, as reproduced in the 4th Edition of Remote Sensing and Image Interpretation, by T.M. Lillesand and R.W. Kiefer:

Graphical explanation of layover and foreshortening.

In topographic feature A, the beam reaches the top of the slope facing the wave front before it reaches the slope base. Thus the upper beam is reflected to the receiving antenna before the lower beam, producing the layover effect. This is most pronounced in near range encounters and decreases as the beam encounter tend towards far range; thus layover has dimininished in feature B. Layover is the phenomenon that results when the radar wavefront reaches the top of a slope before the bottom. Foreshortening begins when the wavefront reaches the bottom before the top, such as where the facing slope is less steep than the impinging wave front. The angle of the opposing face (back side) relative to the look angle produces a third phenomenon known as radar shadowing. Shadow length and darkness increases from feature A through D as the look angle becomes progressively less than the fixed back slope angle.

Another kind of distortion occurs in flat terrain. Geometric figures experience dimensional changes as the Look Direction proceeds from near to far slant range. This is expressed as compression of regularly shaped features (square crop fields may distort into a rhombus). This is evident by the gradual shift from square agricultural fields to rectangular ones in this image, from left to right:

Slant distortion of square fields into rectangular fields.
We can convert these images to ground range images if we know independent information on topography. Still other distortions come from erratic aircraft motions during flight (which are nil from stable space platforms). These can also compensated for these by further processing.

Radar images made from aircraft are more likely to show layover and foreshortening than do images acquired from space radar systems. At higher altitudes, satellites have higher depression angles so that radiation reaches fore and back slopes at about the same time. This also minimizes differences related to near and far range timing.

The foreshortening effects tend to diminish and appear more uniform throughout the image when the radar unit is well above the surface. This is the case for radar on orbiting satellites. The next image displays this diminuted effect well. It's a Seasat image of part of the Pine Mountain thrust in North Carolina:

Seasat Radar image of the Pine Mountain thrust area in western North Carolina.

But, distortion does occur where the geometry of the scene relative to the radar beam brings about some differences in arrival times of the wavefront. Here is layover in part of the Alaska Range as imaged by Seasat:

Layover distortion of steep mountain slopes in the Alaska Range.

Another mode of image distortion, not strictly geometric, is speckling. Speckling produces a granularity to a radar image that can be distracting. This "noise" results from sporadic, almost random, variations caused by small, but recordable, phase shifts in the radar signal which may be related to ground irregularities, such as plowed soil in a field that results in point sources which backscatter in various directions. The speckling can be minimized by resampling methods in image processing. Here is an example of speckling:

Large speckle artifacts in fields imaged by radar.

Effect of Illumination Direction

Linear features in rolling or mountainous terrains, such as long, straight valleys or ridge crests, normally stand out with a combination of bright slope-shadow effects. But their patterns change depending on their orientations relative to the flight line or look direction. We illustrate this effect with an interesting analog experiment conducted by Professor Donald Wise (University of Massachusetts), who used a three-dimensional topographic map covered with dark powder that was illuminated at low artificial light angles from several directions:

Appearance of a three-dimensional topographic map (with contour lines hidden), as illuminated artificially, with light source coming from a N 30 E direction.


Same setup but illumination now coming from a N 60 E direction.

A simple glance at the two images shows an immediately obvious difference: In the top image, ridges and valleys that trend N 60° E are strongly enhanced by lighting coming from the N 30° E direction , whereas these features when illuminated from the N 60° E direction are subdued in expression but those that trend N 30° E (upper part of ridge trend) now stand out. Maximum visual emphasis occurs when illumination is perpendicular to the trend of a linear feature. Other features at varying orientations are visible in each scene but with subdued expression. As we saw on page 2-8, this phenomenon - that linear trends are greatly influenced by illumination azimuth and angle - is quite pronounced in Landsat images, producing a directional bias. That bias is evident from the plots of fractures on azimuthal rose diagrams, where those oriented northeast-southwest (roughly perpendicular to mid-morning sun angles) tend to dominate the distribution of orientations. This trend can be an advantage in airborne radar imagery, because we can chose flight line directions to underscore and accentuate certain directions of interest to optimize detection of fractures in all orientations.

This dependency on direction of illumination is used to good advantage in conducting radar flights over terrains that are sensitive to differences in orientation of mountain ridges, fault valleys, and other conditions where linear features are being sought. The next (aerial) radar image clearly indicates that a different viewpoint (and information extraction) results when the illumination direction shifts relative to the orientation of the scene (north as reference).

Mountaineous area in the Cascade Range (Washington state), seen in two views in which the illumination angle of the radar beam comes from different directions relative to north.
From Drury, S.A., Image Interpretation in Geology, 1987, Allen & Unwin

The effects can be seen by keying in on several features, such as the stream valley at (e), as these "rotate" with the scene shift indicated by the north arrows. This is the same topography but its appearance, as controlled by changing slope illumination, seems quite different in the two view. Note that foreshortening lessens from top to bottom; front illuminated slopes close to the top (near range) display layover.

To bring this point home, consider this image of Precambrian rocks in the African nation of Nigeria.
Radar images of crystalline terrain in Nigeria, with the left image having a look direction from the bottom and the right from the right.

In the left image, the look direction is from the bottom; in the right image illumination is from the right. Same scene - but with notable differences in the expression of topography and structural features.

Source: http://rst.gsfc.nasa.gov