Engineers from the Moscow Aviation Institute have developed a computer vision-based system. It automatically detects hidden defects in composite materials. These materials are used in the production of aviation and rocket technology.
Engineers proposed a new approach to quality control of composite structures, combining the capabilities of X-ray three-dimensional scanning, an electron microscope, and computer vision algorithms. The technology allows for the human-free detection of voids, cracks, and delaminations that can reduce the reliability of aircraft parts.
Previously, finding such damage required manual processing of a large array of images obtained from sample research. Analysis took days and even weeks. The new algorithm performs this work in a few minutes, automatically highlighting potentially dangerous areas of the material.
The methodology consists of several stages. First, the system performs a volumetric analysis of 3D scanning data and identifies internal defects. Then, it examines images of problematic areas obtained using an electron microscope to clarify the size, shape, and exact location of damage at a microscopic level. After that, all information is combined into a single digital model, which helps not only to determine the location of the defect but also to understand the possible causes of its appearance.
According to the developers, the system is capable of recognizing damage smaller than one thousandth of a millimeter. Combining the results of different research methods allows for the most complete picture of the material's condition.
Tests have already been conducted on samples of carbon fiber reinforced plastics and fiberglass used in the manufacture of aircraft elements. The results confirmed the high accuracy and stability of the algorithm's operation.
Earlier, "Pervy Tekhnichesky" reported that the KnAAZ plant and MAI began a systematic analysis of production losses in aviation.