16.20. dg_mosaicΒΆ

This tool can be used when processing Digital Globe images (Section 5). Such an image product may be packaged as several sub-images and corresponding camera XML files. dg_mosaic will mosaic these sub-images into a single file, and create the appropriate combined camera XML file.

The tool needs to be applied to the sub-images which form the left image, and then in the same way to obtain the right image. These can then be passed to parallel_stereo (Section 16.47).

Digital Globe camera files contain, in addition to the original camera models, their RPC approximations (Section 8.19). dg_mosaic outputs both types of combined models. The combined RPC model can be used to mapproject the mosaicked images with the goal of computing stereo from them (Section 6.1.7).

dg_mosaic can also reduce the image resolution while creating the mosaics (with the camera files modified accordingly).

Some older (2009 or earlier) Digital Globe images may exhibit seams upon mosaicking due to inconsistent image and camera information. The --fix-seams switch can be used to rectify this problem. Its effect should be minimal if such inconsistencies are not present.

Digital Globe images can be used as they are as well, but in that case each left sub-image needs to be paired up with one more more right sub-images, create individual DEMs, and then mosaic those into a single output DEM. Hence this program simplifies the data processing.

Only level 1B images are supported (see the product info).


dg_mosaic WV03_20160925183217*P1BS_R*C1-012844055010_01_P001.tif

This will create 012844055010_01_P001.r100.tif and 012844055010_01_P001.r100.xml, where r.100 stands for the full 100% resolution.

Care should be taken to not mosaic together PAN and multispectral images, which have P1BS and M1BS as part of their names.

Command-line options for dg_mosaic:

-h, --help

Display the help message.


Choose the output resolution in meters per pixel on the ground (note that a coarse resolution may result in aliasing).

--reduce-percent <integer (default: 100)>

Render a reduced resolution image and XML based on this percentage. This can result in aliasing artifacts.


Skip RPC model generation.


Skip TIF file generation.

--rpc-penalty-weight <float (default: 0.1)>

The weight to use to penalize higher order RPC coefficients when generating the combined RPC model. Higher penalty weight results in smaller such coefficients.

--output-prefix <name>

The prefix for the output .tif and .xml files.

--cache-size-mb <integer (default = 1024)>

Set the system cache size, in MB, for each process.

--band integer

Which band to use (for multi-spectral images).

--input-nodata-value <float>

Nodata value to use on input; input pixel values less than or equal to this are considered invalid.

--output-nodata-value <float>

Nodata value to use on output.

--ot <Byte|UInt16|Int16|UInt32|Int32|Float32 (default: Float32)>

Output data type. If the output type is a kind of integer, values are rounded and then clamped to the limits of that type.


Fix seams in the output mosaic due to inconsistencies between image and camera data using interest point matching.


Ignore the fact that some of the files to be mosaicked have inconsistent EPH/ATT values. Do this at your own risk.


Render a small 8 bit png of the input for preview.

-n, --dry-run

Make calculations, but just print out the commands.