Nissan Uses AUTIS Inspection System to Slash Paint Defects
The camera-based AI system has reduced defects in paint by 21 percent.
At a Glance
- AI machine learning system finds 7 percent more flaws in half as much time as before.
- The array of 42 cameras spots flaws as small as 0.2 mm.
- Improved detection contributes to a 21 percent reduction in flaws thanks to revised processes.
Human quality inspectors in Nissan’s factories have historically spotted between 85 percent and 95 percent of flaws in the finish of cars exiting the paint booth. That wide range of variability is due to inspector fatigue, reports Nissan new model paint engineer Rod Lynch.
In comparison, the AUTIS AI inspection system is consistently 98 percent to 99 percent accurate.
Technicians using the system can monitor defects that are detected through the overhead display monitor and on wrist-mounted smartphones. AUTIS is good at spotting flaws, but it needs help from the technicians for the AI to learn how to classify these flaws.
“With the feedback from the technician it will be able to identify the signature of those defects, whether it is a piece of dirt or a crater or a fiber and it will just continue to learn and refine and we’ll build a library of what our defects look like,” Lynch said.
Nissan’s immediate response to this information is to have technicians correct the identified paint flaws before the painted body proceeds to the assembly line, which should keep customers happy in the near term. The longer-term benefit, however, is to learn from this information what the causes are of the defects and eliminate them at the source rather than correcting them after the fact.
When blemishes are identified, technicians classify them to help teach the system. NISSAN MOTOR CORP.
“AUTIS has been instrumental in helping us update our cleaning processes to limit defects, improve our robot covers to decrease the amount of lint and fibers introduced to the system, and monitor materials to ensure we don’t see a spike in defects,” said Travis Fritsche, Nissan paint process engineer. “These are just a few items that we’ve improved based on feedback from the system. The system is used daily to monitor shop performance and make improvements. Since launching the system, we’ve seen a 21 percent defect reduction based on changes we’ve made just from the data the system is providing.”
Process issues are categorized by vehicle type and color to target problems that are specific to certain hues or models. "AUTIS can group defects to help pinpoint patterns and improve our processes," Fritsche said.
AUTIS uses 42 cameras that shoot 15,000 images of the painted body to find blemishes in the finish. Nissan’s previous system used lasers that could detect flaws as small as 0.3 mm. This new camera-based system spots defects that are just 0.2 mm and does it in half the time of the laser system while catching 7 percent more of the flaws.
Overhead monitors in the inspection booth tell technicians when paint flaws have been found. NISSAN MOTOR CORP.
Nissan started the project with installations on Smyrna, Tennessee’s paint line one and at Canton, Mississippi’s line two, but it has since expanded to all lines at both of those plants plus the ones at Aguascalientes, Mexico. These facilities build models such as Altima, Frontier, LEAF, Pathfinder, Rogue, Kicks, Sentra, Versa, and Murano.
About the Author
You May Also Like