FEATURE:MANUFACTURING
A revolutionary road: machine vision in automotive manufacturing
As advances in machine vision technology continue to optimise efficiency, flexibility and quality in automotive manufacturing, Lauren Haigh explores the latest developments and how potential challenges can be surmounted to successfully integrate machine vision and drive the industry forward
A
s machine vision has evolved, its adoption in the automotive industry has rocketed, with
possible applications expanding from use in automated inspection to encompass integration in tasks throughout the entire production process. The surge in the use of machine vision is driven by the automotive industry’s need for quality assurance, maximised efficiency and the use of innovation to push boundaries and accelerate growth.
Automation, quality and safety Given that machine vision systems are able to inspect parts faster and more accurately than humans, with the potential for around- the-clock examinations, machine vision continues to play a crucial role in quality inspection in automotive manufacturing. “Machine vision systems excel at
automating complex quality control tasks, identifying defects and ensuring consistent quality,” says Simon Rünzi, Global Industry Manager, SICK Sensor Intelligence (SICK). “For example, foreign object detection detects foreign objects on simple surfaces, but also on demanding surfaces such as high-voltage batteries.” Tom Lambert, Machine Vision Specialist,
EMEA, Zebra Technologies, agrees that, due to the extensive range of materials, components and finished parts that need to be visually inspected for compliance, safety and quality, machine vision is crucial: “Modern machine vision solutions, including those with deep learning software tools, are being used for quality and end of line inspection, traceability of parts, gauging and measurement, presence/ absence checks, metrology and porosity inspection for welding.” Another key focus area is identifying and tracking. “Machine vision helps track parts across
the manufacturing process and verifies that the correct components are being produced,” Rünzi tells Imaging & Machine Vision Europe (IMVE). “For example, the car body can be identified reliably using ultra-high frequency RFID (radio frequency identification) technology at any production step.” Rünzi said that, by identifying the position of target objects, machine vision assists in guiding robots during pick-and- place operations. Lambert agreed that robotic picking is a further significant application: “We see machine vision being used as part of robotic picking, sorting and assembly solutions, where the machine vision camera and software guide the actions of the robotic arm”.
Deep learning for optimal outputs A key advancement in machine vision in automotive manufacturing is deep learning, which facilitates heightened automation, precision and intelligence. Ravishankar Rajagopalan is one of
IMVE’s Visionaries interviewees for 2024 and the director and founder of AUGURAI, which provides vision solutions integrated with optics, AI, automation and robotics. “Recent AI algorithms driven by deep
learning are helping crack use cases which are not feasible with traditional image processing algorithms,” he says. Rünzi agrees that deep learning and
intelligent algorithms are an important advancement. “These algorithms enhance
10 IMAGING AND MACHINE VISION EUROPE AUGUST/SEPTEMBER 2024
defect detection, monitor production parameters, ensure safety and optimise productivity,” he says. Deep learning-based optical character
recognition (OCR) systems can be used to inspect components; for example, reading labels or serial numbers, as well as identifying parts and components and enabling traceability. An additional use is in inspecting surfaces for defects and detecting discrepancies in paint quality. “It [OCR] is a good example of a
new, low/no-code approach to machine vision. It comes ready out of the box with a pre-trained neural network and requires no specialist experience or knowledge,” says Lambert. “It can deliver high-quality, accurate barcode, VIN (vehicle identification number) and serial number reading needed for track and trace, while dealing with stylised fonts, blurred, distorted or obscured characters, reflective surfaces and complex, non-uniform backgrounds.” Such tools are in demand in the automotive manufacturing industry. Zebra’s AI Machine Vision in the Automotive Industry Benchmark Report found that 56% of automotive business leaders surveyed in the UK are using some form of AI in their machine vision projects. In Germany, that figure is 43%. There are also advanced imaging
technologies, such as high-resolution cameras and 3D imaging systems that are enhancing precision, quality and efficiency, as Szymon Chawarski, Product Manager for Vision Solutions, Teledyne DALSA, points out: “3D cameras and scanners are quickly being adopted. 3D cameras using lasers and structured light allow automakers and suppliers to measure parts with complex surfaces. Some surfaces that do not show well using normal camera techniques can also be more easily measured using laser lines or structured light, for example contoured rubber parts or shiny and reflective parts”.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39
