Super-resolution applied to power line inspection with long-range drones
In the SuperPower project (I-NERGY Open Call I), FuVeX has developed AI algorithms to automatically enhance the quality of power lines visual images. The use of this technology has been validated with real operations for an early adopter.
Over 5 million kilometers of European power lines must be inspected regularly. One of the critical data to perform such inspection and detect defaults in these vital infrastructures is high-resolution visual images. Currently, these photos are captured by crewed helicopters which can carry heavy high-resolution sensors (5 mm. per pixel GSD). However, these aircraft are expensive (150 € per kilometer inspected), have personal risks, and are pollutant.
In FuVeX we are working to replace crewed helicopters for our long-range drones in power line inspections eliminating human risks and emissions while lowering the costs up to 94%. One of the main challenges for our drones to perform such inspections is precisely to acquire high-resolution data due to the low-weight of the camera that drones can carry (10 times lighter than the ones carried by helicopters).
To be able to overcome this challenge, FuVeX applied to I-NERGY first open call with the goal of applying super-resolution algorithms to double the number of pixels in images of power lines. To do so, we have trained Generative Adversarial Networks (GAN) obtaining a combination of Convolutional Neural Networks (C-CNN) using thousands of images of real power line inspections. With these photos, we have been able to increase our images GSD up to 2 mm. With this quality imagery, we are able to compete with crewed helicopters thus increasing safety, eliminating human risks and reducing costs. Moreover, having images with up to 2 mm. per pixel allows:
-Better implementation of AI technoogies to fully automate the inspection processes.
-Detect small defects in key elments such in as conductors and cotter pins that cannot be detected with lower resolutions thus improving the safety of the grid.
Moreover, the result of the project increases operational safety as drones can fly higher to obtain good quality images as well as the competitive advantages of the company as the data quality is better than competitors not using super-resolution.
The project results have been validated by UFD - Naturgy group which is the third biggest Spanish utility and has the role of early adopter in this project. As a result of this validation, UFD - Naturgy published the following press release (in Spanish): https://www.naturgy.com/sala_de_prensa/notas_de_prensa/1s2022/naturgy_y…
This project has received funding from the European Union's Horizon 2020 research and innovation programme within the framework of the I-NERGY Project, funded under grant agreement No 101016508