MISB compliant FMV is a rich data source, but is not always a final product in itself.
Given one or more FMV files, a tool should be available to:
- Analyze FMV footprints (from one or more videos) to determine if suitable overlap exists. Optionally, the user can determine a target extent or a target feature for coverage.
- If suitable overlap exists, export required frames required for processing.
- Optionally add ground control/check points.
- Build Orthomosaic/3D Model/Point Cloud at the user's option.
This workflow is straight forward enough without additional tools. However, the power and speed would come from automatically analyzing one video, or an entire catalogue of videos to rapidly determine if the desired output product is possible. Critically, the tool would also determine the best frame export criteria from the video files. For example, in an area with multiple buildings, there might be only one video in a catalogue that can see a particular side of the building.
As far as I can tell, currently there's no way to do this automatically. Analysts have to rely on experience and educated guesses to composite multiple videos. Even in the case of a single video there's no method to determine the minimum necessary frame export to produce a quality 3D model.