Video Metadata/WebVMT: Search Frames By Location

[rjksmith]

Outline
An experiment to test how a video archive containing footage in mixed web formats, i.e. MPEG, WebM and OGG, with metadata (including geolocation) from disparate sources could best be searched by location to return a time-ordered sequence of video frames.

1. Establish demonstrable benefits of using a common metadata format.
2. Determine the merits of non-embedded metadata versus an embedded approach.
3. Refine existing use cases and identify any further applications.

Use Cases

• Remote Maintenance
  Drones can be used to fly around and monitor valuable assets in a remote or inaccessible location, e.g. off-shore wind turbines, to allow low-cost inspection and to help diagnose any problems. Still images form a sequence that can determine the time at which damage occurred to an asset, monitor any subsequent deterioration in its condition, and enable the correct remedial action to be taken at the relevant moment.

  • https://innovateuk.blog.gov.uk/2017/03/15/autonomous-wind-turbine-inspection/

• Flood Monitoring
  Flood levels can be assessed by aggregating video footage from a variety of disparate sources, including drones, helicopters, dashcams, body-worn video, helmet cameras and smartphones, which are regularly stored in an archive. The sequence of still images allows water levels to be determined at different locations over time in order to help predict and monitor flooding in the area.

  • https://www.gov.uk/government/news/exercise-tempest-tests-the-environment-agency-flood-response-ahead-of-winter

Further Details

• WebVMT.org website, including overview, tech demos & blog
• WebVMT Editor's Draft at W3C

[rjksmith]

Area monitoring use case added to WebVMT Editor's Draft