dump-klv

This program displays a subset of metadata derived from the KLV packets embedded in a video file. The default behavior is to print the metadata to standard output for all frames of the video. This can be restricted to a subset of frames. The summarized metadata can be saved to a separate file in either csv or json format, or the full interpreted contents of the KLV packets can be saved to json (-e klv-json).

The default configuration of this tool depends on algorithm implementations in the FFmpeg and KLV arrows which will only available if the KWIVER_ENABLE_FFMPEG and KWIVER_ENABLE_KLV CMake flags are enabled. If you are planning on using the tool to dump frame images, the default configuration depends on the OpenCV so this feature will only be available if the KWIVER_ENABLE_OPENCV CMake flag is enabled. To produce json output the Serialize JSON is needed, so the KWIVER_ENABLE_SERIALIZE_JSON CMake flag should be enabled.

kwiver dump-klv       [options]  video-file

Required arguments:

video-file - name of video file.

Options are:

-h, --help

Display applet usage.

-c, --config filename

Provide configuration file.

-c, --config arg

Configuration file for tool.

-o, --output arg

Dump configuration to file and exit.

-l, --log arg

Log metadata to a file. This requires the JSON serialization plugin. The file is structured as an array of frames where each frame contains an array of metadata packets associated with that frame. Each packet is an array of metadata fields. Alternatively, the configuration file, dump_klv.conf, can be updated to use CSV instead.

--frame-range expr

Frame range to process, indexed from 1. May be a single number or two numbers separated by a hyphen, either of which may be omitted to process from the start or to the end of the video, respectively, e.g. ‘5’, ‘10-100’, or ‘64-‘.

-f, --frames extension

Dump frames into the given image format.

--frames-dir path

Directory in which to dump frames. Defaults to current directory.

-d, --detail

Display a detailed description of the metadata.

-q, --quiet

Do not show metadata. Overrides -d/–detail.

-e, --exporter format

Choose the format of the exported KLV data. Current options are: csv, json, klv-json.

--multithread

Use multithreading to accelerate encoding of frame images. Number of worker threads is not configurable at this time.

--compress

Compress output file. Only available for klv-json.

Default configuration

# Default configuration for core dump-klv applet

video_reader:type = ffmpeg
video_reader:vidl_ffmpeg:time_source = misp
metadata_serializer:type = csv

image_writer:type = ocv