Working With Geospatial Video

Geo-Spatial Video: Understanding and Applying 

Technology for Better Situational Awareness

How valuable would it be if you could take video from a moving remote source, such as a UAS platform, and get accurate GPS location information for what that video is showing. Even more, how valuable would that information be if it was in real time or Near Real Time (NRT). A user could then monitor the NRT video streaming from an active drone and see what the drone is looking at and where it is located. This information would be invaluable to users such as law enforcement, that is actively pursing a suspect, or a wild fire control manager that is monitoring the location and direction of a forest fire and making decisions for control measures based on that Full Motion Video (FMV). Full motion video, also known as Geospatial video or “mobile mapping”, is a technology that merges GPS coordinate and location information with video (Mills, 2010). There is a difference between FMV and geospatial Video as FMV usually refers to NRT video and geospatial video has been recorded and processed.

Geospatial video is highly useful and in many regards a required part of UAS operations for firefighting, civil and military application and anywhere else video to map correlation is required. Other drone users would not benefit from the large meta data sets that FMV require. The movie industry for example, would not need to know the location as much as producing an aesthetically pleasing vantage point. For the applications where geospatial video is more useful than “normal” video, it is a dramatically improved resource. Watching video feed, especially UAS video, it is easy to get “tunnel vision” and loose all sense of direction or orientation. Its to easy for a drones camera to be zoomed in to a specific target and have the user loose all sense of position and orientation of the drone itself. An argument could be made that drone accidents have been cause by the operator loosing situational awareness due to focusing on the camera feed and not the aircraft. Having metadata that produces coordinates of what the camera is focused on and presenting that information on a digital map would provide the ability to keep the drone in LOS with the target while maintaining a safe operation.

One other example of how FMV video could be better than traditional video is that while using alternative sensors, IR, FLIR, multi-spectral etc., it may not be entirely clear what the camera is focused on. Having that information would increase accuracy as the target could be verified using GPS metadata.

Using Open-source Software to Create Geo-spatial Video

The following example of geospatial video was created using “GeoTagger Free”, an open source software application from Remote GeoSystems.  The idea was to take a video with accompanying GPS data points and the merge the two so that the location of the images could be known.  
The first step in creating a map to video product was to upload the data and video into the geo-tagging software. 

GeoTagger reference
 point selection

 Once the video and data was in the program, then the software must have a reference point to base the other data points off for the most accurate position information. Ground Control Points (GCP's) would be very beneficial for this part of the process. In absence of a formal GCP, predominate land features can be used as  GCP's for correlation of the video to the data. Once the location and video has been "linked" the software will process the video to show the location of where the image was taken on the map with a cross-hair and follow along while the video plays.  With the file loaded, the path will be colored green, after the reference point is selected the path will be orange, and when the software as process the data, the path will be color coded red ( As depicted in the following images)

Reference point selection based on video imput.. 

Software adjustments point for accuracy of points.

Final product creation with target
showing location that of video

Additional features can bee added with the purchase of the Pro version such as the metadata of the platform, altitude, heading and speed. In the free version, the only feature in the additional information window is the time.                 
The product produced is a guide that assigns data points to timestamps of the video file. It can be easily seen that this type of correlation would not be the high standards of accuracy needed for survey type of applications. This software produces a product that acts more of a guide and not exact representation of the video to the ground. Regardless, this type of information is accurate enough for many types of applications. Looking at large objects to get a general orientation like when planning or  analyzing the layout of vendors for a community event, for example.  Or searching for a moving object, this Geo spatial video could provide a location.  The video used for this demonstration can be found here: GeoTag Video.

Additional features of software

References

Mills, Jacquelin W. (2010). Geospatial video for field data collection. Applied Geography. Vol 30. Issue 4. Pg 533-547. Retrieved from https://www.sciencedirect.com/science/article/abs/pii/S0143622810000342