Automobile seat covers presence/absence inspection
Confirm that covers have been installed on automotive seats
Graphical programming environment for deep learning-based industrial image analysis
With exterior design constrained by safety and aerodynamic requirements, automakers have put increasing attention on their interiors, providing a luxurious appearance even in economy models. One important differentiator is the upholstery of seat covers, which come in a wide variety of materials, patterns, and textures.
As seats are manufactured, seat covers are automatically placed over them. Verifying the presence of a seat cover before the seat is passed to the next phase can be surprisingly difficult. If a seat cover is missing, the seat must be pulled out of the line and returned.
Seat covers are made out of soft and flexible materials and can differ significantly in positioning and arrangement while remaining acceptable. Because of the wide range of appearances acceptable seat covers may have, conventional machine vision is not reliable in determining presence and absence.
Cognex Deep Learning easily verifies the presence of a seat cover. It trains on small sets of images of both seats with installed seat covers in a range of missing, good, and bad installations. The classification tool learns to divide all images into either present or absent conditions with high accuracy. If a new seat cover design is introduced, operators on the line can easily train the tool to detect the presence or absence of the new seat cover type.
Quick presence/absence verification detects any problems immediately and minimizes delays on the line.