Peachtree City, Ga. — Artificial intelligence is helping manufacturers eliminate defects. And it's rapidly getting better, according to Scott Heins, director of sales and marketing for Intravis Inc.
"What I really see in the future is AI will be the main way to do vision inspection, and professional tools will be a last resort," Heins said. "It's really a matter of days."
Heins spoke about AI and vision systems for packaging at the Society of Plastics Engineers Annual Blow Molding Conference, held Oct. 7-9 in Peachtree City. He explained that automation and sustainability revolutions are underway in the plastics industry, with inspection playing a major role as an enabler.
The expanded use of AI with improved connectivity and closed-loop controls is helping container makers accurately inspect products at high speeds.
With traditional vision systems, users have to set up rules and criteria for the equipment to use to process and decide whether containers pass inspection.
With AI, companies are teaching the system what a defective product looks like.
"If you have a manufacturing process that is constantly changing, adaptive learning is the way to go," Heins said. "There can be a means to do continuous feedback."
Detection false positives and false negatives can be improved with the use of AI and giving more knowledge to the system.
Another way AI can be used is providing data that allows a company to monitor the quality of products. Companies then can use this comprehensive data to be proactive about improving any problems.
"AI tools live and breathe on massive amounts of data. Data is the real future of manufacturing," Heins said.
By using a centralized data base, like the cloud, companies can access real-time data regardless of physical location. This can allow for maximizing production and efficiency within a company by allowing employees to access this data and respond to what is being shown.
"For us, it's more than just a buzzword," Heins said. "Intravis has been using AI for many, many years. Today it's really a choice of choosing AI when it's a better solution."
The ways that Intravis has been incorporating AI include for precise position detection of preforms. This allows for a consistent and accurate preform and reliability in monitoring.
The company also uses AI to boost sleeve label inspection accuracy. AI can understand the difference between defects and harmless deviations.
"It's a pretty challenging application," Heins said. "It's inspected 360 degrees all around," he said.
With the labels sometimes being shiny, having bright spots or having water spots of moisture left over, the AI can look past those characteristics and be able to product less false positives.
Addressing the beginning of the AI movement, Heins acknowledged that there were many random mistakes being made by the software, but it has continued to learn and has a better knowledge base now.