Leonardo has chosen Reco 3.26 to improve the inspection process of materials on aerostructures.
The following are the solutions offered by Reco:
With a view to building a flexible tool capable of automatically detecting and classifying defects in different production cycles, Reco 3.26 is developing an industrial research project and experimental development of artificial intelligence techniques for improvement and automation. non-destructive and fiber testing during blemishes resulting from ultrasonic testing in carbon fuselage sections. The is to provide NDI inspectors with a tool capable of automatically acquiring potentials without the need for the inspector to analyze all of the scan maps.
The visual inspection is carried out with the aid of optical microscopy in order to acquire the photos on which the qualification operators will perform the measurements for assessing the severity of any wrinkles detected.
As part of the structural analysis for the identification of these specific anomalies, Reco has developed a prototype machine vision system, for the detection and measurement of wrinkles, in fuselage sections composed of carbon fibers. Reco WrinkleMeter is a software for processing images acquired by microscope by quality control operators, to identify and measure the information characterizing any wrinkled ply inside the package.
Aluminum Material Ispection
As part of the structural analysis for the identification of anomalies in the fuselage assembly process, Reco 3.26 is developing an experimental computer vision system which, using suitably trained convolutional neural networks (CNN Convolutional Neural Network), is able to detect, locate and classify possible defects on aluminum surfaces caused by improper use of assembly tools or abrasive media and / or incorrect rework operations.
The computer vision subsystem developed by Reco 3.26 and enhanced with artificial intelligence techniques and algorithms, receives in input the images and / or information retrieved from the acquisition subsystem, and returns the classification of the defect, if present, together with its location on the image.