Our solutions include computer vision algorithms for automatic processing and analysis of the images. This means that the inspections allow systematic and repeatable measurements, detection of distresses and identification of faults. Computer vision minimises human error and improves efficiency. This technology is applied to all of our visual solutions for PAPI, ALS, aerodrome lighting and PCI inspections.
In the example below, an ALS is automatically detected from the rest of elements in the airport and then it can be analysed to determine the number of bars and the lights per bar. With that an automatic diagnostic can be provided identifying missing lights or other faults.
The use of Convolutional Neural Networks (CNN) and machine learning techniques allow generalisation and add robustness, being able to detect a wider range of faults in different situations, weather and light conditions.
In the example below, our database is fed with thousands of real runway distresses (according to the norm ASTM D5340). The system is then able to identify distresses in the images captured by the drone and provide a detailed report thanks to our convolutional neural networks.
We also use deep learning techniques such as convolutional neural networks to identify and classify the paved areas of the airport (runway, taxiway and apron), so that the system can detect the different distresses that are included in the norms for each type of pavement (asphalt y PCC).
These techniques allow the system not only to detect and classify the distresses, but also help in the accurate location of the distresses so that the complete reports can be elaborated and analysed by areas, as specified by the norm.
This technology can be updated and adapted to many other applications to improve inspections & maintenance inside the airport and even in other fields, such as civil engineering or construction.