Multi-Surface Inspection and Classification with COI
Inspect multiple surfaces for defects and then classify them
While most manufacturing processes in the electronics industry are now automated, some defects are ambiguous and can only be seen at certain angles. Frequent production line changes make developing a new vision application extremely challenging. That means inspecting defects on an assembled consumer electronics product like a mobile device still requires manual inspection. Additionally, human inspection results vary from one inspector to another and struggle to scale with the quantity of cosmetic defect inspections that is required in the consumer electronics industry.
Available in Asia only, Cognex’s Cosmetic Optical Inspection (COI) system is designed to capture multiple, high-detailed images of electronic device surfaces using a vision guided robotics system. The robot presents the part to a single high-resolution camera to capture surface images under ideal lighting conditions at different angles. Deep learning software then analyzes those images to identify defects that might have been missed if the part was only inspected from one angle. Defects are then classified to provide business intelligence data on the manufacturing process. This removes the guesswork to identify defects accurately and consistently at scale before the part is passed along the production process.