The AI-based machine learning system can see the material's internal structure by looking at it from the outside. The system does not use things like X- or gamma-rays. And that means it does not use dangerous radiation. If this kind of system can tell the concrete element's condition from regular camera images that thing makes damage searching more effective. Developers of that system used the metal pipe as an example. But there is possible to use that system also in other types of materials.
"A machine-learning method developed at MIT detects internal structures, voids, and cracks inside a material, based on data about the material’s surface. On the top left cube, the missing fields are represented as a gray box. Researchers then leverage an AI model to fill in the blank (center). Then, the geometries of composite microstructures are identified based on the complete field maps using another AI model (bottom right). Credit: Jose-Luis Olivares/MIT and the researchers". (ScitechDaily.com/MIT’s AI System Reveals Internal Structure of Materials From Surface Observations)
"One potential application of the new method is nondestructive testing; you no longer have to open a metal pipe, for instance, to detect defects. Credit: Courtesy of the researchers" (ScitechDaily.com/MIT’s AI System Reveals Internal Structure of Materials From Surface Observations)
The scratch with a certain type of bottom outside the metal tells that in its structure is something wrong. In the same way, things like fungus or some other flora tell that there can be internal damage to concrete elements or bricks.
If the system can get a hint about internal damage by using regular images it can call X-ray systems to take a closer look sooner than before. In this version, machine learning means that the system can connect X-ray images to regular images. And then it can make a database of what certain types of damages look like from the outside. If there is suspected damage, the system can call the X- or gamma-ray system for a closer look.
https://scitechdaily.com/mits-ai-system-reveals-internal-structure-of-materials-from-surface-observations/
Comments
Post a Comment