The system taps a neural network to register problems associated with welding and pauses production so these components can be corrected.
Agtech capabilities are bringing traditional farming into the 21st century. These solutions range from sprawling LED-equipped indoor farming facilities to robotically plucking ripe produce off the vine using computer vision and artificial intelligence (AI). On Thursday, John Deere and Intel announced a pilot program that relies on AI and computer vision to detect defects in manufacturing related to the welding process.
“Welding is a complicated process. This AI solution has the potential to help us produce our high-quality machines more efficiently than before,” said Andy Benko, quality director at John Deere Construction and Forestry Division. “The introduction of new technology into manufacturing is opening up new opportunities and changing the way we think about some processes that haven’t changed in years.”
Manufacturing and welding challenges
John Deere uses Gas Metal Arc Welding at 52 of its global factories and this process involves “hundreds of robotic arms” fed millions of pounds of weld wire every year, the company explained, and a common production challenge relates to porosity in which trapped gas bubbles cause metal cavities, decreasing weld strength.
SEE: TechRepublic Premium editorial calendar: IT policies, checklists, toolkits, and research for download (TechRepublic Premium)
Historically, manual processes involving skilled techs have been used to detect such welding flaws, although such attempts “to deal with weld porosity issues during the welding process haven’t always been successful,” the release said, and defects identified later in the production pipeline lead to disruptions which require reprocessing or “scrapping of full assemblies.”
AI and computer vision defect detection
The John Deere and Intel pilot program leveraged an end-to-end software and hardware solution to “generate insights” in real time “at levels beyond the human sense’s capability,” the release said. This system taps a neural network-based inference engine to register manufacturing defects and then pauses the welding process in real-time, allowing the company to correct these flaws as needed.
SEE: Natural language processing: A cheat sheet (TechRepublic)
“Deere is leveraging AI and machine vision to solve a common challenge with robotic welding,” said Christine Boles, vice president in Intel’s IoT Group and general manager of Industrial Solutions Group. “By leveraging Intel technology and smart infrastructure in their factories, Deere is positioning themselves well to capitalize not only on this welding solution, but potentially others that emerge as part of their broader Industry 4.0 transformation.”
Nuts and bolts: System hardware and software
This AI-enabled defect detection system uses Intel i7 processors, Intel Movidius VPUs, the Intel Distribution of OpenVINO toolkit alongside an industrial-grade ADLink Machine Vision Platform as well as a MeltTools welding camera.