Manufacturers in various sectors face challenges in hiring and retaining a qualified workforce, keeping up with technological innovation, and meeting growing expectations for speed and precision.
Sustainability, digitization of operations and supply chains, increased expectations for safety and demand for personalization are also top priorities for those in the automotive manufacturing industry.
According to Zebra’s recent Automotive Ecosystem Vision study, 73% of industry decision makers believe their business will be at a competitive disadvantage if they do not adopt more digital technologies, and they believe that “developing software expertise is listed in the top five investment priorities. A recent report from McKinsey agreed, saying that with recent developments in robotics, AI and machine learning, we are on the cusp of a new era of automation.
One of the key innovations of increasing importance is deep learning optical character recognition (OCR) software. A survey of original equipment manufacturer (OEM) decision makers surveyed said industrial machine vision usage will increase by 83% by 2027, according to Zebra’s automotive research. ~44%) is expected to increase.
The combination of deep learning and OCR meets the need for speed, accuracy, and reliable solutions for compliance, quality, and presence checking across the manufacturing industry. Operational leaders in the automotive, pharmaceutical, electronics, and food and beverage industries will now be able to take full advantage of deep learning OCR.
smart manufacturing
Machine vision and deep learning OCR enable smart manufacturing, which Gartner defines as coordinating physical and digital processes within factories and across other supply chain functions. They transform how people, processes and technology operate to provide the information needed to impact the quality, efficiency, cost and agility of decision making. In other words, driving automation with deep learning and OCR gets the most out of your hardware, software, and people.
However, getting the OCR check right can be difficult. Stylized fonts, blurry, distorted, or unclear characters, reflective surfaces, or complex uneven backgrounds make it impossible to achieve consistent results using traditional OCR techniques there is.
However, new tools are emerging on the market that provide industrial-quality deep learning OCR and have ready-to-use neural networks pre-trained with thousands of different image samples. This new series is capable of delivering a high level of accuracy out of the box, even when dealing with the most difficult cases.
At the automotive manufacturing site, a deep learning OCR solution can accurately read printed, embossed, matte, and metallic serial numbers stamped on batteries, tires, parts, and accessories so that they can identify the correct vehicle identification number (VIN). ), which means that you can check that it matches. These solutions can accommodate a wide variety of font styles and sizes, as well as changing “challenging” lighting and manufacturing environments.
Deep learning OCR can also be used as part of a broader machine vision solution. For example, on the automotive manufacturing floor, machine vision solutions can be deployed to inspect connector pin inspection, PCB conformal coating, adhesive inspection, wire harnesses, batteries and polarity, general assembly presence, quality and compliance. increase. inspection.
In these scenarios, a machine vision solution deployed using the same machine vision camera and integrated software platform inspects items much faster, flagging suspect defects and errors to engineers so they can be identified as defects. You can check and determine if there is and move on.
Feeding review decisions back to the neural network keeps continuous learning input active to further develop and enhance the model. This optimizes efficiency and removes important but tedious manual work from engineers.
Growing Value of Deep Learning
The speed and accuracy of deep learning can greatly assist engineers in ensuring manufacturing quality, controlling production costs, and improving customer satisfaction. However, ease of use is equally valuable, and that is the strength of deep learning OCR software. This is an easy-to-implement and easy-to-use application that requires no machine vision expertise and can be deployed in a few simple steps.
The combination of more accessible machine vision and deep learning OCR solutions opens up new possibilities for industrial imaging professionals and engineers to think and act like data scientists. This development is both necessary and welcome as the speed, volume and variety of data continues to grow and expectations for higher levels of speed, security and accuracy are high.
can reach myself or wider machine vision team for further discussion.
About the author
Rudolf Schambeck is Senior Channel and Market Development Manager for Machine Vision at Zebra Technologies, Germany, where he has been with the company since 2021. Prior to joining Zebra, he worked at Cognex, a machine he vision company, for five years, sales he was an engineer, business development, and account he was a manager. His role focused on B2B and automotive. Prior to joining Cognex, he worked at Intercontec Produkt GmbH and at Irlbacher Blickpunkt Glas GmbH where he held B2B and OEM sales engineer, account he manager and sales roles.
