Cosmetic Housing Inspection
Search for specific defects, such as scratches, while tolerating unimportant anomalies and variations on device housing
Graphical programming environment for deep learning-based industrial image analysis
Cosmetic inspections can be challenging when parts vary, whether at the component level or at the packaging and housing level. Scratches, dents, and other cosmetic defects may not affect functionality but do affect finished quality and consumer perception. Some cosmetic defects may be obvious cause for rejection, while other minor defects are acceptable. For this reason, manufacturers need to train an inspection system to search for specific defects and differentiate them from minor blemishes. Programming an inspection of this complexity into a rule-based algorithm requires complex defect libraries. Human inspection, though more flexible, is too slow, unreliable, and inconsistent.
Using the defect detection tool in supervised mode, an engineer can train Cognex Deep Learning to search for specific defects, such as scratches, while tolerating unimportant anomalies and variations. The tool is optimized to work with images that are low-contrast or are poorly captured. For example, the image below illustrates how the defect detection tool analyzes both good and bad images of earbuds. During runtime, the software characterizes images with severe scratches as defective, having learned to recognize and ignore minor cosmetic blemishes.