Image (Re-)projections and Merging - Open Source Geospatial Tools: Applications in Earth Observation - page 107

Fig. 8.4 Pixels that are not
filledinthewarpingprocess
can be set as no-data ( dark
colored ) so these can be
ignored for subsequent
processing
-cblend distance
Set a blend distance to use to blend over cutlines (in pixels).
-crop_to_cutline
(from GDAL 1.8.0) Crop the extent of the target dataset to the extent of the
cutline.
-overwrite
(from GDAL 1.8.0) Overwrite the target dataset if it already exists.
-nomd
(fromGDAL 1.10.0) Do not copy metadata. Without this option, dataset and band
metadata (as well as some band information) will be copied from the first source
dataset. Items that differ between source datasets will be set to * (see -cvmd
option).
-cvmd meta_conflict_value
(from GDAL 1.10.0) Value to set metadata items that conflict between source
datasets (default is “*”). Use ”” to remove conflicting items.
-setci
(from GDAL 1.10.0) Set the color interpretation of the bands of the target dataset
from the source dataset.
srcfile
The source file name(s).
dstfile
The destination file name.
8.3.1 Reproject Images
When (re-)projecting an image, it is likely that some areas in the target image cannot
be filled with valid data since the bounding box is altered during the transition.
Consequently there are typically gaps produced in the target image (see Fig. 8.4 ).
Gaps in a reprojected image can be labeled as “no-data”. Subsequent image
processing algorithms can treat these values accordingly. As an example, when
calculating the image average value, these values should be ignored. In addition, an
alpha channel can be assigned to the output file, which makes the no-data transparent.