init-cameras-landmarks

The init-cameras-landmarks tool estimate cameras and landmarks from a set of feature tracks.

The default configuration of this tool depends on algorithm implementations in the FFmpeg, MVG, and OpenCV arrows which will only available if the KWIVER_ENABLE_FFMPEG, KWIVER_ENABLE_MVG, and KWIVER_ENABLE_OPENCV CMake flags are enabled.

kwiver init-cameras-landmarks       [options]

Options are:

-h, --help
Display applet usage information.

-c, --config arg
Configuration file for tool

-o, --output-config arg
Output a configuration. This may be seeded with a configuration file from -c/–config.

-v, --video arg
Input video

-t, --tracks arg
Input tracks

-k, --camera arg
Output directory for cameras

-l, --landmarks arg
Output landmarks file

-g, --geo-origin arg
Output geographic origin file

Default configuration

# Default configuration for the track-features applet

# If true, all mask images will be inverted after loading.
# This is useful if mask images read in use positive
# values to indicated masked areas instead of non-masked areas.
invert_masks = false

# If true, extract metadata from the video and use metadata to
# guide which frames are tracked.
use_video_metadata = true

# A majority of the time, mask images are a single channel,
# however it is feasibly possible that certain
# implementations may use multi-channel masks. If this is
# true we will expect multiple-channel mask images,
# warning when a single-channel mask is provided. If this
# is false we error upon seeing a multi-channel mask image.
expect_multichannel_masks = false

# select the descriptor type
descriptor_type := ocv_ORB
descriptor_binary := true

# Parameters for the feature tracker
block feature_tracker
  include core_feature_tracker.conf

  # The maximum number of frames for the GUI to process.
  # The tracker will choose frames distributed over the video
  max_frames = 500
endblock

block output_homography_generator
  include ocv_homography_estimator.conf
endblock