16.45. parallel_bundle_adjust

The parallel_bundle_adjust program is a modification of bundle_adjust (Section 16.5) designed to distribute some of the preprocessing steps over multiple processes and multiple computing nodes. It uses GNU Parallel to manage the jobs in the same manner as parallel_stereo. For information on how to set up and use the node list see Section 16.47.

The parallel_bundle_adjust tool has three processing steps: statistics, matching, and optimization. Only the first two steps can be done in parallel, and in fact after you have run steps 0 and 1 in a folder with parallel_bundle_adjust, you could just call regular bundle_adjust to complete processing in the folder. Steps 0 and 1 produce the *-stats.tif and *.match files that are used in the last step.

In particular, if you would like to repeat bundle adjustment, regular bundle_adjust would suffice as well, since the supporting data are already created. If you would like to do some other experiments with the cameras or the options of the tools, in a different output folder, you can first copy there the existing content of an existing run, and as before, call bundle_adjust, rather than its parallel version. (Also consider using --force-reuse-match-files and --skip-matching, to prevent bundle_adjust to try again to compute match files if the cameras are too new or if some match file computation failed before and if it is likely to fail again if re-attempted.)

The match files created by this tool can be used by bundle_adjust and parallel_stereo via the options --match-files-prefix and --clean-match-files-prefix.

Command-line options for parallel_bundle_adjust:

--nodes-list <filename>

The list of computing nodes, one per line. If not provided, run on the local machine.

-e, --entry-point <integer (default: 0)>

Bundle adjustment entry point (start at this stage). Options: statistics = 0, matching = 1, optimization = 2.

--stop-point <integer(default: 3)>

Bundle adjustment stop point (stop before this stage). Options: statistics = 0, matching = 1, optimization = 2, all = 3.

--verbose

Display the commands being executed.

--processes <integer>

The number of processes to use per node. The default is to use as many processes as cores.

--threads <integer>

The number of threads to use.

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

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

-v, --version

Display the version of software.

-h, --help

Display the help message.