16
ERDC TR-05-1
3
IMAGE POST-PROCESSING
3.1
CAMIS
The same post-processing algorithms were used in both the Poplar Island and
the Sny levee projects. These techniques are used to correct both radiometric and
geometric image distortions. The steps include:
Band-to-band registration (a geometric correction).
Removal of frame edge darkening and frame center hot spots (a
radiometric correction).
Frame-to-frame mosaicking (a geometric correction).
Mosaic geometric registration (a geometric correction).
3.1.1
Poplar Island
Degradation of the radiometric (or color balance) quality of digital imagery is
the result of influences from a variety of external and internal parameters. Ex-
ternal parameters include: solar azimuth angle, solar zenith angle, atmospheric
conditions (e.g., aerosols, water vapor), and surface bi-directional reflectance
factors. Typical internal parameters that degrade image radiometric quality in-
clude lens distortions, lens field darkening, CCD anomalies, and system opera-
tion error (e.g., f-stop settings, shutter speed).
Software developed at TEC radiometrically corrected the complete set of
CAMIS images. The first step was to band register each frame. Because the four
cameras are aligned in sequence, with each lens sited parallel to the others, each
band of a single frame has its center, or principal point, offset from the other
bands. Figure 8 shows a single false color CAMIS frame before and after band-
to-band registration.
The next post-processing step applied a single algorithm to address several of
the external and internal parameters that influence radiometric fidelity. Figure 9
displays the visual impact of the correction algorithm on a true color composite.
Typical radiometric distortions appear as a general darkening towards the edges
of the frame and a bright area (hot spot) at the center. The same brightness shifts
are present within most aerial photography, particularly the pronounced hot spot,
or glaring, at the center of the photo.