Drones are getting smarter – so smart, in fact, that our drones are now self-learning. This means that our drones can recognize different objects, parts, or damage independently. This process can be set up with a low threshold, so that your company can come into contact with deep learning software.
Deep learning software is a software that we can tailor make. The software can start learning for itself. A number of examples that the software can learn are: recognizing rust damage, recognizing series numbers, recognizing certain parts and giving reports of serious damage that must be replaced immediately.
Initially this will all be new to you. An example of the software’s accuracy is that is able to recognize the difference between Skittles and M&Ms. This is even difficult to recognize with the naked eye, but the software can see exactly which is an M&M and which is a Skittle. We use the same principle for various types of applications, such as train tracks, roads, bridges, aircraft and much more.
The biggest advantage is the accuracy. The software sees more than people can see with the naked eye. We have already experienced this in various scenarios, where for example the inspector overlooked things, but the drone indicated that there was clear damage.
Another big advantage is the time saved. A screen full of inspection photos is a thing of the past. The self-learning software removes good, bad and doubtful cases from the data, so that the inspector still largely controls the inspection.