A team of researchers from Austria recently developed an AI-powered drone that can track moving objects through dense foliage. So much for escaping to the forests and living off the land if the machines ever rebel against us.
According to the teams research paper†
While detecting and tracking moving targets through foliage is difficult (and often impossible) in ordinary aerial photos or videos, it becomes practically achievable with image integration…
This finding, together with the implementation of a first drone-powered camera array for parallel synthetic aerial aperture imaging, makes it possible to present the first results on tracking moving people through dense forest. In addition to people, other targets (e.g. vehicles or animals) can be detected and tracked in the same way.
This can affect many application areas, such as search and rescue, surveillance, border control and wildlife observation.
In front: There are numerous ways for an interested party to track moving objects through dense foliage, including FLIR and other thermal optical systems, but this offers a new AI-powered ripple: the ability to track color changes through occlusion.
Teaching an AI to figure out what happens in an image when a significant amount of relevant information is hidden is one of the biggest challenges of the AI world. In this case, the researchers developed a system capable of using a technique called “color aberration detection” in real time to track hidden objects in motion — a first, according to the diary of remote sensing†
Background: The team created a lightweight 1D camera array that captures aerial images of high-foliage areas via drone deployment. Traditional aerial photography techniques use color aberration detection to scan images for clusters of pixels that do not match the natural environment.
However, the team’s contribution was to develop a combination of hardware and machine learning that combined traditional techniques with their new aerial camera system to create something that can identify color aberrations and track them through closed areas.
Quick take: This could have immediate and potentially huge benefits for search and rescue efforts. Due to its relatively low computer requirements and power overhead, this is a solution that can be shipped relatively quickly to remote parts of the world and put to use immediately. I see this saving the lives of hikers who have drifted off course or plane crash survivors who are trapped in remote areas.
It could also have a huge benefit for conservation efforts if used in conjunction with GPS tagging and other tracking efforts.
But it’s also clear that something like that could turn a Predator drone, for example, into something that can track targets through closed-off areas (such as congested city blocks or jungles), even if it loses connection to its human controllers or its communications freeze.