Vricon research initiative with U.S. Special Forces delivers 30cm detail on-demand
Just last month, we released a new video describing an interesting project we performed for the Special Operations Command Foreign Comparative Test (FCT) program. Check it out here. The project’s goal was to deliver a custom 3D mapping suite of an area unavailable for drone-based observation and then determine the value of the results.
Basically, FCT asked us to demonstrate how fast we could generate our best quality 3D product (30 centimeter resolution) over a large geographic area. The FCT clock would start ticking when they made the specific request and stop when we delivered the data to the end user.
The video explains in detail the product we delivered (in only 27 hours!) to FCT, and when I reflect on the video summary, a number of interesting points occur to me. My thoughts run from “how do we best show and deliver such data” to “what does it really mean to have data with accuracy better than a meter?”
So, let’s explore some of this in this blog. While our product for FCT is similar to our baseline product, it differs on a few critical points worth noting:
- Density. The FCT product it is more “dense.” This makes sense, because density is the number of bits of information needed to hold the data. Our regular products are VERY lightweight, computationally speaking – only 2 to 5 megabytes per square kilometer, depending on the format. This means that an area the size of Mississippi fits on a mini-SD card like the one you use for your camera. When the data is taken at 30cm resolution, the density jumps by a factor of 4. This makes it similar to the density of a point cloud – but the information content is much greater (5-10 times on average).
- Accuracy. Because the resolution of each pixel (that’s “voxel” in 3D talk) is one third of a meter instead of one half, accuracy—the ability to correctly locate the voxel in space—improves greatly. So far, all indications are that the system can deliver a sub-1-meter 3D position on each element observed. This is a big deal! Why? Because instead of being able to locate a room in a building, you can locate a specific window, door, or desk. (This assumes you can see into the area, of course – we are still looking at visible solid objects after all.) Both the potential for precision strike and material or object identification improve significantly. For example, think of the difference between identifying a chair with a person in it versus an empty one. At one meter, the color information can actually tell the difference.
- Sharpness. In the FCT product, objects are sharper and the “street view” is more reliable. Standard Vricon products, while highly accurate at all levels of “zoom”, have a street-level look that is what we call “melty” – like an ice cream cone beginning to suffer the sun. The resolution for the FCT product resulted in a very obvious reduction in melt, thus increasing the usability of the information.
These are all arguments for covering more areas at higher resolution. The current constraint is that there is only one satellite providing the source imagery at the improved level – Worldview 3. As Maxar deploys the Legion Constellation next year, the area of Earth captured at higher resolution will potentially increase quite rapidly.
Like any good GEOINT geek, I admit to being very excited about the potential, and our way forward. Till next time,