Plastic Injection Molded Part Inspection
Inspect small plastic parts for structural and cosmetic defects
Even small plastic parts have complex surfaces and indentations that need to fit precisely with other parts. Sink marks, warping, short shots, and flash are all defects that affect the functionality of plastic parts.
Other defects are purely cosmetic, not affecting function but having a significant negative effect on sales if they are in a visible part of a consumer product. These cosmetic defects include scratches, flow lines, burn marks, dimples, and discoloration.
Catching these defects upstream prevents them from impacting later assembly steps, which can require the scrapping of a fully assembled piece. The difficulty of lighting these small parts in the constrained spaces where they emerge from the mold, combined with their reflective, wavy, and complex surfaces, makes it difficult to accurately inspect plastic molded parts with conventional machine vision, particularly at production line speeds.
Cognex Deep Learning is ideal for cosmetic visual inspection of subcomponent assemblies. The defect detection tool trains on a set of images of a range of defect-free plastic parts. The defect detection tool then identifies any anomalous parts, detecting both functional and cosmetic defects that differ from the trained images.
If the component is invisible after final assembly, making cosmetic defects such as small color variations acceptable, the defect detection tool trains with those variations as part of the training set, and will then not mark them as anomalies. If defect definitions change as final assemblies and subsequent sales are analyzed, the defect detection tool retrains quickly with another set of images with a different acceptability range.