“Having artificial intelligence technology onboard accelerates our roadmap to full automation of the drone workflow.”
Aerialtronics, in conjunction with Neurala, a pioneer in deep learning software, and NVIDIA, has demonstrated automated inspections by an intelligent drone with deep learning capabilities at the GPU Technology Conference Europe.
Aerialtronics and Neurala collaborated to make the demonstration on the Altura Zenith UAS, which incorporates the NVIDIA Jetson TX1 module. The resulting system can visually inspect a cell tower and recognize the equipment mounted on the mast. This is the first step required to start automating the documentation of assets, and assessing the mechanical functionality and condition of the cell tower to identify rust and other defects.
The intelligent drone application is expected to be extended beyond cell towers to include inspection of bridges, buildings, wind turbines and other infrastructure as business benefits from scalability and rapid adoption of new technology.
The real-time processing of the data stream is made possible by combining the Aerialtronics unmanned aerial system and smart dual camera payload with Neurala’s deep learning neural network software, which is capable of finding and recognizing objects in flight, using the NVIDIA Jetson TX1 platform.
“Not only can we do clever things with the vision and thermal data, we can connect the flight computer so the drone is fully aware of it surroundings,” said Robin van Putte, chief of product strategy at Aerialtronics. “We don’t want to end our inspection mission with a SD card in our hand and invest a lot of time in data offload and post-processing. Instead we want to be able to immediately verify the results and take quick decisions. Having artificial intelligence technology onboard accelerates our roadmap to full automation of the drone workflow.”
“Inspections of cell towers, buildings, bridges and more by intelligent drone will transform industries. Neurala and Aerialtronics working together are taking drone autonomy to the next level,” explained Massimiliano Versace, CEO of Neurala.
This is a tremendous combination of market leaders.
NVIDIA was founded in 1993. Located in Silicon Valley, they are best known for their high performance graphics processing units (GPUs) for the gaming market, as well as system on a chip (SOC) units for the mobile computing and automotive market. Less celebrated but no less awesome are their high speed computers specifically built for AI. Their Jetson platform is described as “the embedded platform for autonomous everything.” Which should just about cover it. Their boards are also use by Kespry.
Neurala, located in Boston, describes itself as “The brains behind the smartest devices.” Drones are one of the verticals they are targeting. “Neurala can classify what it sees as normal or as containing a certain class of abnormality. A probability statistic enables the user to determine the sensitivity of the system.” They partner with NVIDIA.
Aerialtronics is a Dutch company that is well known for its UAV aircraft, sensors and software which are used by T-Mobile among others. In September 2016, Aerialtronics announced a relationship with IBM Watson to “connect Watson’s visual recognition application programming interfaces, IBM’s cloud, drones and high-definition cameras to perform inspections on assets such as wind turbines, oil rigs, and cell towers. Clearly, the company sees a tremendous opportunity here.
This goes far beyond autonomous flight operation. This is about getting more and more value from the inspection process so that there is less back-end human interpretation. As we have seen with drone military ops, it is the tasking and analysis that really chews up the human resources. Whether a person launches the drone and recovers it or not, these are the first steps in developing systems that will prioritize the need for human intervention. Very exciting.