Publications

Version: 1.0 Approved

LDO-ID 2020-09-10-04

Title Automatic Composite Parts Defect Detector

Abstract

Both Computer Vision and Artificial Intelligence are acquiring an important role in the industrial processes, complementing human work to aid and improve the production of goods. In the aeronautic (and especially aerostructures) field, for example, it is crucial to determine as soon as possible whether a part of CFRP (Carbon Fiber Reinforced Plastic) is acceptable or not. Such a check is currently performed by hand at Aerostructure Division by qualified personnel who carefully analyse many parts images. This process is not only time consuming, but it can also introduce delays in the production process, causing great economic losses. To complement and relieve the human work, here an automatic method is proposed that performs image analysis to separate key carbon fibre layers from background, and performs defect detection by computing specific figures on the extracted components. Results are made available to the personnel, who validate the final output.

Authors Corvaglia Stefano, Gallo Nicola, Baecchi Claudio, Del Bimbo Alberto, D’Attis Giovanni, Nobile Giorgio

Type Paper for Seminar/Symposium/Conference

Media EU FP7 Framework Programme CPS4EU - Cyber Physical Systems for Europe (H2020-ECSEL-2018-IA - Grant Agreement n°: 826276, https://cps4eu.eu/)

Web site

https://cps4eu.eu/downloads/publications/

Anno 2021

Cancel