Fernerkundung I (Digitale Bildverarbeitung) - Friedrich-Schiller ...

spatially by up to 1/2 a pixel, causing a jagged or blocky appearance if there is much
rotation or scale change.
Bilinear Interpolation:
Bilinear interpolation determines the grey level from the weighted average of the four
closest pixels to the specified input coordinates, and assigns that value to the output
coordinates. This method generates an image of smoother appearance that nearest
neighbor, but the grey level values are altered in the process, resulting in blurring or loss
of image resolution. Because of these changes in the grey level values, any image
classification processes should be performed before the interpolation. Bilinear
interpolation requires 3 to 4 times the computation time of the nearest neighbor method.
Cubic Convolution:
Cubic convolution determines the grey level from the weighted average of the 16 closest
pixels to the specified input coordinates, and assigns that value to the output
coordinates. This method is closer to the sin(x)/x resampler than nearest neighbor or
bilinear interpolation. The image is slightly sharper than that produced by bilinear
interpolation, and it does not have the disjointed appearance produced by nearest
neighbour interpolation. Because the grey level values are altered by this method, any
image classification processes should be performed before the interpolation. Cubic
convolution requires about 10 times the computation time required by the nearest
neighbor method.
Sin(x)/x 8pt and 16pt:
8pt determines the grey level from the weighted average of the 64 closest pixels to the
specified input coordinates and assigns the value to the output coordinates. 16pt does
the same, using the 256 closest pixels. The image is sharper than that produced by
bilinear interpolation and it does not have the disjointed appearance produced by nearest
neighbor interpolation. Because the grey level values are altered by this method, any
image classification processes should be performed before the interpolation. Sin(x)/x
with an 8 x 8 window requires about 20 to 40 times the computation time required by
the nearest neighbor method. Sin(x)/x with a 16 x 16 window requires 40 to 80 times
the computation time required by the nearest neighbor method.
Mosaicking:
Mosaicking is the blending together of several arbitrarily shaped images to form one
large, radiometrically balanced image, so that the boundaries between the original
images are not easily seen. Any number of geocoded images can be blended together
along user-specified cut lines (polygons).
Mosaicking is a special case of Geometric Correction where the registration takes place
into an already existing image. If GCPs are collected, the uncorrected image is
transformed according to the derived polynomial, into the georeferenced image. If no
Mosaicking GCPs are provided, but both images already have compatible georeferencing,
then an appropriate translation and scaling will be applied instead of a polynomial
transformation (note: this usually does not lead to a perfect fit of image borders and will
create map-border-faults).
Mosaic Cut Line:
When images to be mosaicked overlap, it is useful for the user to specify a cut line
determining which image takes precedence in different areas. This is done in GCPWorks
on the Mosaic Area Collection panel. The cut line is a polygon, enclosing the region to be
replaced in the georeferenced image. Any of theimage pixels within the polygon will be
replaced by pixels from the uncorrected image if possible. Georeferenced pixels will not
be replaced if the corresponding location is of the uncorrected image, or if the pixel value
on the uncorrected image is the Background Value.