AI Can Speed Car Seat Quality Inspection by More Than 27 Times

Greater accuracy in under 3 seconds of inspection time

Published: Thursday, September 21, 2023 - 11:01

(EasyFlow: Kaunas, Lithuania) -- Artificial intelligence can significantly increase accuracy and reduce time for car seat quality inspection. EasyFlow’s EasyODM—a dedicated AI and computer vision software—was deployed at a production facility of a leading global car seat manufacturer. Over the course of two years, the software was able to reduce car seat inspection time to 2.2 seconds, with an overall performance accuracy of 99%.

With area scan cameras installed on the production line, the system was able to identify such manufacturing defects as wrinkles, scratches, staining, and creases. Defects could be attributed to a certain seat area and—depending on the manufacturing workflow—defective seats could be forwarded for automated robotic or manual defect removal.

“Manual seat inspection usually takes around 60 seconds,” says Simas Jokubauskas CEO of EasyFlow, creator of EasyODM. “By utilizing AI, this process can be reduced to only 2.2 seconds on average. Inspections by humans for simple defects or low-volume production, while accurate, remain monotonous, time-consuming, and susceptible to fatigue-caused errors. AI introduces uniformity in quality control, operating at and even exceeding human levels of accuracy.” 

A premium car brand can offer more than 40 versions of seats across different product lines, but EasyODM can easily adapt to support different seat models, layouts, trims, and finishes. Legacy car-seat quality inspection software was developed to work only with a particular seat model, meaning that it needed to be calibrated to support new products. It was also not tolerant of manufacturing realities—for instance, that a seat for inspection could arrive slightly bent or a few centimeters off from the ideal inspection point.

EasyODM was developed to accommodate a diverse product portfolio and often imperfect manufacturing environment. At first the software identifies the seat in the image and isolates it from the background noise. Then, pixel by pixel, the software scans segments of the seat for defects, highlighting the affected regions in the image. The highlighted defects can then be forwarded for automated removal. After defects are eliminated, the software performs an additional quality check.

EasyODM was extensively tested at the facilities of one of the largest car seat manufacturers. With more than 250 manufacturing and assembly plants globally, the company provides services to leading automotive brands.

“EasyODM software has a transformative value for automotive manufacturers,” says Jokubauskas. “Its killer feature—the ability to support multiple seat designs with little-to-no extra effort—provides tremendous ROI for end clients. In a competitive, fast-paced automotive market, this allows suppliers to quickly adjust to new trends and requirements, giving them a significant competitive edge.”

EasyFlow develops AI applications for retail, construction, and manufacturing industries.

About The Author

Quality Digest

For 40 years Quality Digest has been the go-to source for all things quality. Our newsletter, Quality Digest, shares expert commentary and relevant industry resources to assist our readers in their quest for continuous improvement. Our website includes every column and article from the newsletter since May 2009 as well as back issues of Quality Digest magazine to August 1995. We are committed to promoting a view wherein quality is not a niche, but an integral part of every phase of manufacturing and services.