The SensiCut system will automatically detect the material before laser cutting

Currently being developed in the department Computer Science and Artificial Intelligence Laboratory At MIT, the SensiCut system may finally go to the laser cutters currently sold to minimize the risk of confusion on production lines. The currently used solutions based on optical technology are much less accurate, and already in the design phase SensiCut is as much as 98% effective.

The SensiCut system has been designed to prevent starting cutting with incorrect power or speed settings, depending on the material used. It does this in a simple (at least on paper) way, identifying 30 different materials by their surface, and itself consists of a laser pointer, an image sensor, a Raspberry Pi Zero microprocessor and a rechargeable battery. All this is in a simple 3D-printed casing.

Also read: 485,681 graphics cards were returned to the owner after a Chinese court ruling

MIT scientists are working on the SensiCut system to improve every laser cutter and make life easier for operators

In practice, SensiCut is to be integrated with laser cutters. As soon as a given material is placed on them, it is to start examining its unique microstructure thanks to the differently reflected laser light, which is picked up by the built-in image sensor. The collected information is then transferred to a computer, where the system based on a neural network is able to match this pattern to one of the known materials.

Also read: The Danish Ordrupgaard Museum has a new underground gallery

Based on this, the computer will provide the operator with data on the type of material and, if it decides that it can be processed, will indicate the ideal power and speed settings for the cutting laser. At this point, SensiCut can even laser scan the entire surface of a flat multi-material object, determining which areas are made of which substances to better guide the laser during operation and automatically adjust laser power and speed.

Want to stay up to date with WhatsNext? Follow us on Google News