The American company Gecko Robotics, known for its inspection robots for industrial facilities, has introduced a new Stratosight system, a drone for autonomous roof inspection. Now, instead of risking the lives of workers by sending them to slippery roofs, enterprises can entrust this task to a flying robot. Although, to be honest, it's hard to imagine that checking the roofs was such a serious problem that it required the creation of a specialized drone.

Stratosight is positioned as a solution for large commercial and industrial buildings where regular roof inspection is a necessity. The drone automatically circles the roof, creating a 3D map and detecting damage using a thermal imager and high-resolution cameras. The manufacturer claims that the system can detect problems invisible to the human eye, apparently including those mythical leaks that roofers have been unable to find for years.

The "autonomy" of the system is particularly admired. The drone allegedly plans the route itself, avoids obstacles and lands to recharge. Although, given the sad experience with autonomous drones in other areas, it can be assumed that in practice there will still be an operator with a remote control somewhere nearby, just in case.

Interestingly, Gecko Robotics originally specialized in robots that crawl through pipes and tanks. Apparently, they decided that the next logical step was to move from horizontal surfaces to inclined ones. Although, if you think about it, roof inspection is not the most obvious evolution for a company called Gecko.

The economic feasibility of the system raises questions. The cost of the drone and the software clearly exceeds the cost of several years of work by ordinary roofers. But the manufacturer assures that this will pay off by preventing serious damage — they say it's better to pay for a drone today than for a new roof tomorrow. Although, perhaps, it would be easier to just regularly carry out high-quality repairs.

Critics may point out that the drone will not be able to replace a person completely — for example, it will not take a sample of the material or tap on the surface to identify voids. But the developers claim that their machine vision technologies can handle these tasks remotely. I would like to see how the algorithm determines the degree of waterproofing wear based on the photo.

The funniest aspect of this story is the seriousness with which the company treats its development. They really believe that the world has been waiting for an autonomous roof inspector, as if there are no other, more pressing problems in robotics.

However, it cannot be denied that the system has potential, especially for high—rise buildings or facilities with dangerous roofs. The main thing is that drones do not start signing repair contracts on their own, otherwise they will have to create roofer robots to fulfill their recommendations.