Ramses Valvekens (Easics) & Daan Gheysens (Robovision)
AOI of PCBs using deep learning on FPGA
Automated optical inspection (AOI) of printed circuit boards (PCBs) is a critical step in assuring the quality of assembled boards. Pinpointing defects in this step and correcting them is far better than doing so later on, during electrical testing, or worse in the field. The large variety of electronic components makes it hard to keep AOI software up to date. Deep learning is therefore ideally suited for this task. We will demonstrate how we do this with FPGA technology, using a scalable solution. It takes advantage of ever-increasing FPGA computational power and keeps up with component diversity into the future.