CLI Parameters

CLI Parameters

Default pipeline: aliceVision_cameraInit -> aliceVision_featureExtraction -> aliceVision_imageMatching -> aliceVision_featureMatching -> aliceVision_incrementalSfM -> aliceVision_prepareDenseScene -> aliceVision_depthMapEstimation -> aliceVision_depthMapFiltering -> aliceVision_meshing -> aliceVision_depthMapFiltering -> aliceVision_texturing

aliceVision_cameraInit

Required parameters:

-i [ --input ] arg

A SfMData file (*.sfm) [if specified,--imageFolder cannot be used].

--imageFolder arg

Input images folder [if specified, --input cannot be used].

-s [ --sensorDatabase ] arg

Camera sensor width database path.

-o [ --output ] arg (=cameraInit.sfm)

Output file path for the new SfMData file

Optional parameters:

--defaultFocalLengthPix arg (=-1)

Focal length in pixels. (or '-1' to unset)

--defaultFieldOfView arg (=-1)

Empirical value for the field of view in degree. (or '-1' to unset)

--defaultIntrinsic arg

Intrinsics Kmatrix "f;0;ppx;0;f;ppy;0;0;1".

--defaultCameraModel arg

Camera model type (pinhole, radial1, radial3, brown, fisheye4, fisheye1).

--groupCameraFallback arg (=folder)

When there is no serial number in the image metadata, we cannot know if the images come from the same camera. This is problematic for grouping images sharing the same internal camera settings and we have to decide on a fallback strategy:

• global: all images may come from a single device (make/model/focal will still be a differentiator).
• folder: different folders will be considered as different devices
• image: consider that each image has different internal camera parameters

--allowSingleView arg (=0)

Allow the program to process a single view.

Warning: if a single view is process, the output file can't be use in many other programs.

Log parameters:

-v [ --verboseLevel ] arg (=info)

verbosity level (fatal, error, warning, info, debug, trace).

aliceVision_cameraLocalization

Note

This program takes as input a media (image, image sequence, video) and a database (vocabulary tree, 3D scene data) and returns for each frame a pose estimation for the camera.

Required input parameters:

--sfmdata arg

The sfm_data.json kind of file generated by AliceVision.

--mediafile arg

The folder path or the filename for the media to track

Options for the output of the localizer:

-h [ --help ]

--visualDebug arg

If a folder is provided it enables visual debug and saves all the debugging info in that folder

--outputAlembic arg (=trackedcameras.abc)

Filename for the SfMData export file (where camera poses will be stored).

Default : trackedcameras.abc.

--outputJSON arg

Filename for the localization results (raw data) as .json

Common optional parameters for the localizer:

--descriptorPath arg

Folder containing the descriptors for all the images (ie the *.desc.)

--matchDescTypes arg (=sift)

The describer types to use for the matching

--preset arg (=normal)

Preset for the feature extractor when localizing a new image {LOW,MEDIUM,NORMAL,HIGH,ULTRA}

--resectionEstimator arg (=acransac)

The type of *sac framework to use for resection (acransac, loransac)

--matchingEstimator arg (=acransac)

The type of *sac framework to use for matching (acransac, loransac)

--calibration arg

Calibration file

--refineIntrinsics arg

Enable/Disable camera intrinsics refinement for each localized image

--reprojectionError arg (=4)

Maximum reprojection error (in pixels) allowed for resectioning. If set to 0 it lets the ACRansac select an optimal value.

Parameters specific for the vocabulary tree-based localizer:

--nbImageMatch arg (=4)

[voctree] Number of images to retrieve in database

--maxResults arg (=10)

[voctree] For algorithm AllResults, it stops the image matching when this number of matched images is reached. If 0 it is ignored.

--commonviews arg (=3)

[voctree] Number of minimum images in which a point must be seen to be used in cluster tracking

--voctree arg

[voctree] Filename for the vocabulary tree

--voctreeWeights arg

[voctree] Filename for the vocabulary tree weights

--algorithm arg (=AllResults)

[voctree] Algorithm type: FirstBest, AllResults

--matchingError arg (=4)

[voctree] Maximum matching error (in pixels) allowed for image matching with geometric verification. If set to 0 it lets the ACRansac select an optimal value.

--nbFrameBufferMatching arg (=10)

[voctree] Number of previous frame of the sequence to use for matching (0 = Disable)

--robustMatching arg (=1)

[voctree] Enable/Disable the robust matching between query and database images, all putative matches will be considered.

Parameters specific for final (optional) bundle adjustment optimization of the sequence:

--globalBundle arg

[bundle adjustment] If --refineIntrinsics is not set, this option allows to run a final global bundle adjustment to refine the scene

--noDistortion arg

[bundle adjustment] It does not take into account distortion during the BA, it consider the distortion coefficients all equal to 0

--noBArefineIntrinsics arg

[bundle adjustment] It does not refine intrinsics during BA

--minPointVisibility arg (=0)

[bundle adjustment] Minimum number of observation that a point must have in order to be considered for bundle adjustment