μNET Inspector 1.6

Self-learning neural network

Haiku Tech introduces a vision inspection machine with an outstanding performance/speed ratio. Self-learning AI software in combination with high quality components allows new microscopic scale full object inspections at speeds previously impossible.

The machine comes with a linear motor X-Y vacuum table that transports sheets or plates to the high magnifi­cation line-scan cameras. The inspection information is processed by an intelligent AI server within the 20 ms range per image. The machine can run automatically in-line with production equipment by addition of product handling robots.

The state of the art AI software achieves sophisticated standards of reliability and repeatability at a high inspection speed. The machine‘s self-learning algorithm HiQtensor 1.6 achieves low inspection failure rates and teaches itself how to detect the defects of the future. A user interface displays the location of defects for corrective actions. A microscopic on-screen 2nd opinion check by operators or engineers is available in the software. The smart system allows to increase the yield and quality output of your factory.

• Self-learning neural network
• High speed: < 10 seconds for a 1,6 Gigapixel picture
• Flexible vacuum table holds multiple objects: sheets, hard plate
• Fast change over: No lens changes for different object sizes
• Line scan cameras with 3,5 micrometer pixel size
• Top notch artificial intelligence platform with > 90 TFlops
• Automation available for in-line application