identification and delineation of homogeneous areas with particular surface
properties.
The texture of a SAR scene may be classified as being smooth, fine, linear,
gravity, speckled, flecked or checkered. Various degrees of coarseness can be
applied in the different categories. Factors influencing the appearance of texture
on a radar image include the dynamic range of the SAR system, processing and the
amount of enlargement of the photographic product. These affect the coarseness of
texture and must be considered in the interpretation. The texture on radar imagery is
not as easily affected by the “gain setting” of the SAR system as tone and, therefore,
remains relatively constant from one image to another.
Three components of texture, i.e. micro-texture, meso-texture and macrotexture, can be identified. Micro-scale texture, or speckle, is an inherent characteristics of SAR imagery, because of the sampling statistics of the fading signal,
resolution, and independent sampling, i.e. single or multi-look processing. Even
though a surface may be very smooth, it will still display a very fine, random texture
at the micro-scale level. Meso-scale texture is produced by spatial in homogeneities, usually on the order of several resolution cells. It tends to be spatially and not
randomly organized and relates, for example, to the distribution of vegetation
associations; however, individual elements within a plant community may not be
identified. Macro-scale texture allows the identification and delineation of unit
areas that are located within boundaries of relative homogeneity. These may consist
of an assembly of well-resolved terrain elements, but with different backscattering
levels. In very general terms, macro-sacle texture can also be conceived as a
pattern, if it refers to an orderly spatial arrangement of features on an image.
Patterns can be assessed and described using such categories as orientation, spacing, density or uniformity.
14.7.3 Size and Shape
The size of features in a radar image is essentially a function of the radar system
parameters, the scaling ratio of the imagery and the amount of magnification. High
resolution SAR imagery can be enlarged as much as ten times without suffering an
appreciable loss of detail. Radar imagery are generally smaller in scale than aerial
photographs. Therefore, the images of particular features are smaller than their
equivalents on air photos obtained from similar platforms altitudes. However, SAR
does have an advantage in that its resolution is independent of range, which makes
its rendition of small sized features in far range superior to aerial photography.
Shape is defined as the spatial form of an object or area with respect to a
relatively constant contour, or background. Regular geometric shapes, and the
regular spatial arrangement of certain patterns for that matter, are valuable clues
for distinguishing natural and cultural features. Even though natural features tend to
have irregular outlines, many can be identified by their size and shape alone, for
instance volcanic cones, glacial features or alluvial fans. The shape of natural
14 Digital Processing of SAR Data and Image Analysis Techniques
297
properties.
The texture of a SAR scene may be classified as being smooth, fine, linear,
gravity, speckled, flecked or checkered. Various degrees of coarseness can be
applied in the different categories. Factors influencing the appearance of texture
on a radar image include the dynamic range of the SAR system, processing and the
amount of enlargement of the photographic product. These affect the coarseness of
texture and must be considered in the interpretation. The texture on radar imagery is
not as easily affected by the “gain setting” of the SAR system as tone and, therefore,
remains relatively constant from one image to another.
Three components of texture, i.e. micro-texture, meso-texture and macrotexture, can be identified. Micro-scale texture, or speckle, is an inherent characteristics of SAR imagery, because of the sampling statistics of the fading signal,
resolution, and independent sampling, i.e. single or multi-look processing. Even
though a surface may be very smooth, it will still display a very fine, random texture
at the micro-scale level. Meso-scale texture is produced by spatial in homogeneities, usually on the order of several resolution cells. It tends to be spatially and not
randomly organized and relates, for example, to the distribution of vegetation
associations; however, individual elements within a plant community may not be
identified. Macro-scale texture allows the identification and delineation of unit
areas that are located within boundaries of relative homogeneity. These may consist
of an assembly of well-resolved terrain elements, but with different backscattering
levels. In very general terms, macro-sacle texture can also be conceived as a
pattern, if it refers to an orderly spatial arrangement of features on an image.
Patterns can be assessed and described using such categories as orientation, spacing, density or uniformity.
14.7.3 Size and Shape
The size of features in a radar image is essentially a function of the radar system
parameters, the scaling ratio of the imagery and the amount of magnification. High
resolution SAR imagery can be enlarged as much as ten times without suffering an
appreciable loss of detail. Radar imagery are generally smaller in scale than aerial
photographs. Therefore, the images of particular features are smaller than their
equivalents on air photos obtained from similar platforms altitudes. However, SAR
does have an advantage in that its resolution is independent of range, which makes
its rendition of small sized features in far range superior to aerial photography.
Shape is defined as the spatial form of an object or area with respect to a
relatively constant contour, or background. Regular geometric shapes, and the
regular spatial arrangement of certain patterns for that matter, are valuable clues
for distinguishing natural and cultural features. Even though natural features tend to
have irregular outlines, many can be identified by their size and shape alone, for
instance volcanic cones, glacial features or alluvial fans. The shape of natural
14 Digital Processing of SAR Data and Image Analysis Techniques
297