Track shipments

This drone can track targets through forests

A team of Austrian researchers recently developed an AI-powered drone that can track moving objects through dense foliage. Might as well escape into the woods and live off the land if the machines ever came up against us.

According to the team’s research paper:

Although detecting and tracking moving targets through foliage is difficult (and often impossible) in regular aerial imagery or video, it becomes practically doable with the integration of imagery…

This discovery, associated with the implementation of a first network of cameras piloted by drone for parallel synthetic aperture aerial imagery, makes it possible to present the first results on the tracking of people in motion through the dense forest. Besides people, other targets (eg vehicles or animals) can be detected and tracked in the same way.

This could impact many application areas, such as search and rescue, surveillance, border control and wildlife observation.

In the front: There are myriad ways for an interested party to track moving objects through dense foliage, including FLIR and other thermal optics systems, but this offers a new AI-powered wrinkle: the ability to track color changes by occlusion.

Humanoid greetings

Subscribe now for a weekly recap of our favorite AI stories

Teaching an AI to understand what is happening in an image when a significant amount of relevant information is hidden is one of the biggest challenges in the world of AI. In this case, the researchers developed a system capable of using a technique called “color anomaly detection” in real time to track moving occulted objects – a first, according to the Journal 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 anomaly detection to analyze 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 airborne camera system to create something capable of identifying color anomalies and tracking them at through the occluded areas.

An image of a drone mounted with an array of cameras next to a breakdown of the array's rooms.
Credit: Nathan, et al.

Quick take: This could have immediate and potentially huge benefits for search and rescue efforts. Due to its relatively low computational requirements and power overhead, it is a solution that could be shipped to remote parts of the world relatively quickly and put into use immediately. I can see it saving the lives of derailed hikers or plane crash survivors stranded in remote areas.

It could also have a huge benefit for conservation efforts if used in combination with GPS tagging and other tracking efforts.

But it’s also obvious that something like this could turn a Predator drone, for example, into something capable of tracking targets through occluded areas (such as congested city blocks or jungles) even if it loses the connectivity with its human controllers or if its communications are blocked.