Optical Character Recognition on Wafer Carrier Rings
Accurately read damaged identification codes for reliable traceability
Graphical programming environment for deep learning-based industrial image analysis
Once the wafer is diced, the wafer ID is no longer usable. To keep traceability of the die that was previously created on the wafer, a carrier ring marked with an identification number carries diced silicon wafers until they are removed from the ring for wire bonding. The dicing process spreads shards from sawing all over the die and ring so they must be cleaned. Repeated cleanings degrade the carrier ring’s surface which decreases the code’s readability. Variations in the surface and characters make it hard for rule-based vision technology to accurately read these codes over time. Characters like 0 and O or l and 1 are hard to distinguish if faded or worn. Unreadable rings can cause a slow-down in the automation process, which affects production throughput. Using OCR to read codes on wafer rings enables them to be used longer and keeps the automation process moving.
Cognex Deep Learning tools enable manufacturers to accurately read identification codes on wafer carrier rings, even after they have degraded from multiple cleanings. A smart camera and deep learning software work together to decipher damaged codes using optical character recognition (OCR). The Deep Learning Read tool within the software works right out of the box, dramatically reducing development time, thanks to the deep learning pre-trained font library. Users simply define the region of interest and set the character size. In situations where new characters are introduced, this robust tool can be retrained, without vision expertise, to read application-specific codes that traditional OCR tools are not able to decode.