Pouch Surface Inspection
Cognex Deep Learning inspects pouch-style EV batteries
A pouch-type battery cell is shaped somewhat irregularly during the degassing process. After jig formation, the cell pouch is pressed to iron out and smooth the surface. It is essential that pouches are even, unwrinkled, and unbent. Cell battery manufacturers employ automated inspection systems between these stages to catch any surface defects. The pouch’s complex surface texture creates a noisy and confusing background which can obscure wrinkles, bubbles, and other defects. The visual appearance of one cell pouch can vary drastically from another, making it too complicated and time-consuming to explicitly search for all defects.
Cognex Deep Learning uses deep learning-based vision algorithms to identify defects, such as bubbles and wrinkles, by learning from annotated images. The model learns the normal appearance of a pouch’s surface, including natural variations that don’t constitute defects. All features that deviate from the model’s normal appearance are characterized as anomalous. In this way, Cognex Deep Learning reliably and consistently detects all anomalies without the need for extensive defect libraries.