Smart camera provides instant control for laser welding
A novel camera system has 25,000 processors integrated into its pixels, allowing it to analyse images the instant they are taken and automatically adjust the output of a welding laser.
Developed by Germany's Fraunhofer Institute, the camera is part of a control system for welding car doors. These are usually assembled from overlapping sheets of metal which are welded together by a laser beam which melts them in a spot measuring several-tenths of a millimetre, producing a so-called full penetration hole that closes again when the laser beam moves on.
It is vital for the laser output power to be set correctly. If it is too low, the weld is weak because it does not extend over the full cross-section of the metal sheets. If it is too high, the laser cuts right through instead of simply melting.
Until now, welders have gauged the right laser output by trial and error and then kept it constant. However, if the protective glass gets dirty, hours can pass before it is noticed and during that time the metal sheets may not be properly welded.
To date, it has not been possible to monitor laser welding automatically because closed loop control requires a frame rate of more than 10 kilohertz, or 10,000 images per second.
Now though, researchers at the Fraunhofer Institute for Physical Measurement Techniques (IPM) in Freiburg have developed a control system capable of analysing images over ten times faster than conventional computer software.
"Our system analyses 14,000 images per second and uses the acquired data to adjust the laser output," explained IPM project manager Andreas Blug.
"We use special cameras in which a tiny processor is integrated in each pixel. All these processors - 25,000 in total - work simultaneously. In conventional image processing systems the data are handled consecutively by just a small number of computer processors," he added.
The new systems are called Cellular Neural Networks (CNNs). Just a few microseconds after the image is taken, the camera delivers an analysed picture of the contours of the full penetration hole. For small holes the system increases the output, and for large ones it reduces it.
"In developing this adjustment system we have achieved the first industrial application of CNN technology," said Blug.
A prototype already exists, and the researchers now intend to test the system in production.