Placeholder Machine vision system applied to Automobile Manufacturers | SINSMART

For the manufacturing industry, there are not only complex processes in terms of design, implementation, and optimization, but also need to face strict evaluations from multiple parties, from channel partners and consumers to industry and government regulators. At the same time, production remains sluggish even as the economy begins to recover, so now is an opportune time to take a closer look at and optimize processes and their supporting systems, especially those that determine the traceability of parts and goods, drive product Inspection or quality control, processes and support systems that help reduce waste or affect overall production and order fulfillment rates. And when we took a hard look at the performance of machine vision and fixed industrial scanning solutions in the field of automobile manufacturing, we began to notice that the application of machine vision was not only a new model, but also represented a whole new opportunity.

A new vision for machine vision and stationary industrial scanning solutions from an automaker's perspective

According to the latest data released by the China Association of Automobile Manufacturers, China's automobile production and sales will complete 27.021 million and 26.864 million vehicles in 2022, a year-on-year increase of 3.4% and 2.1%. The total automobile production and sales have ranked first in the world for 14 consecutive years. For the Chinese market, which has a huge automobile market, it is very important to ensure the safety performance of automobiles in every link of the manufacturing process.

More than 30,000 parts are needed to make an average car, and each one has to be impeccably designed and assembled, not just to keep drivers happy, but to keep them safe. If a washer is missing, or a gear is missing a tooth, anything can go wrong when the vehicle hits the road. Just like the supply chain, every vehicle is only as good as every component and system is built to exacting specifications. But it's no easy task for factory workers on a moving production line checking that everything from doors and dashboards to powertrains and pistons are complete and even installed correctly. In addition, without the aid of a high-resolution camera, it is difficult for the human eye to accurately determine the correct application and amount of sealants and adhesives, or to identify design flaws in tiny chipsets, beads or bolts.

By taking advantage of modern machine vision systems, automakers are able to conduct rigorous analysis of parts and products in limited time. If the cameras and software are functioning properly, and operators are able to maximize their capabilities without a long learning curve, it becomes easier to inspect all parts passing by on the production line and make immediate and confident pass/fail decisions. It is very easy because the process is fully automated and requires little operator intervention.

However, many traditional machine vision systems do not operate seamlessly, simply and efficiently, as required by modern production lines. Some low-resolution cameras often make manufacturers question the accuracy of machine vision systems and their ability to consistently find defects on the production line. Or the image processing speed of the old system is too slow to keep up with the pace of today's car production, let alone the requirements of the factory of the future.

The industry generally agrees that while machine vision is a must in today's Industry 4.0 environment, it creates more work than manufacturers can (or want to). Manufacturers want a system that is easy to set up, learn, maintain, and even retrofit as needed to support a variety of applications. The ideal is to have a cross-platform industrial automation solution that can be easily configured to support any number of vision or barcode reading applications. Essentially, manufacturers want to automate part verification, confirm label placement during the quality control stage, and more with just one integrated machine vision and scanning system, such as Zebra's all-in-one machine vision and stationary industrial scanning system correct, and track inventory movement on production lines without having to install multiple solution components across facilities.

Today, a single device can do multiple jobs, including capturing high-quality images, scanning multiple barcodes, and generating process control indicators, making it easier to confirm in real time where parts are, whether components are assembled correctly, and who the manufacturer or Whether the supplier fully complies with the different requirements in the contract, regulations and other quality control measures. In short, the ideal industrial automation solution for automotive manufacturing is finally moving towards reality.

What new possibilities exist when manufacturers only need to integrate, configure, learn and maintain one industrial automation platform?

From an operational or technical standpoint, the four things most automakers focus on are complexity, precision, reliability, and speed. Regarding the application of machine vision in the manufacturing process, it is necessary to deeply explore how new industrial automation solutions can deal with the following four challenges:

Challenge 1: In terms of different production processes and system management, some of them rely on fixed industrial scanners, machine vision cameras or a combination of both. Automakers need to streamline not only these systems, but processes such as parts infeed, quality control, and even outbound logistics.

Solution considerations: Having a platform with both barcode reading and machine vision functions makes it easy to transform and optimize the process at the pace of the manufacturer, allowing the solution to focus on every challenge facing the automaker. It is even more ideal if the platform uses a standardized communication protocol, supports 1D and 2D codes, and can provide a large machine vision library. Automakers need only scrutinize the solution's ability to decode, ensure the credibility of the data generated and stored by the system, and ensure that digital records can be automatically created, whether for images or inventory movements, to help fill out shipping manifests, Inventory management systems, even invoices, etc. This eliminates another (often manual) step in the production, storage and shipping workflow.

Challenge 2: The automobile manufacturing industry has a very low tolerance for assembly errors, and parts can only be classified as compliant or not. It is also important to check the incoming and outgoing dates and batch numbers of each component, package, and pallet.

Solution considerations: Unless you see it yourself, it's difficult to appreciate the level of detail required to evaluate each component, and in many cases these evaluations need to be done in a split second. Zebra has been working with an automaker to refine its machine vision solution for inspecting engine, radiator and housing components. Its engine must be equipped with three bolt caps to pass inspection and enter the packaging process. If even one is missing, or the bolt cap is in the wrong place, the assembly will be reworked. For larger parts, there may be someone who can easily perform a visual inspection, but the speed of inspection is often limited. The engine is used in production vehicles and millions of these components are inspected each year. Considering that bolt caps are one of tens of thousands of items that must be inspected, machine vision cameras, when properly optimized, can be an advanced solution to this type of problem.

In addition, each of the automaker's shell assemblies has 11 bolts that must be rigorously inspected before being delivered to drivers. Therefore, machine vision cameras need to correctly verify the presence and design compliance of each bolt with exceptional precision. If a non-compliant component accidentally passes inspection, damage can occur when the driver tries to tighten the bolt into place. Therefore, it is essential to have advanced algorithms such as pattern recognition, edge detection and blob analysis to ensure that the product meets the requirements.

From a labeling or marking standpoint, each part will have a direct part mark (DPM) and bulk items will have additional labeling mechanisms, all with rich items and statuses. Additionally, getting a good picture of moving objects requires a thorough overhaul of machine vision and stationary industrial scanners to ensure they can read all barcodes for part verification. They should also have OCR tools that can use system information to verify dated lot codes for parts compliance and accurate reconciliation with warehouse management systems (WMS).

Challenge 3: Parts that are out of compliance, or the wrong item received or shipped can have negative consequences. Therefore, our goal is to solve the problem early to avoid a recall, financial loss, or worse.

Solution considerations: Because industrial automation solutions, such as those offered by Zebra, capture and evaluate barcode data and images, defects or process inconsistencies can be easily identified and alleged non-compliance difficult to contest. In the case of non-compliant parts, direct part identification or other identifiers can be used to trace suppliers or line systems so that problems can be investigated and resolved immediately. Stored images can be retrieved for reference if required. In the long run, this increases uptime and reduces scrap. If it's an assembly issue, the team can be retrained on demand. If it is a labeling issue, it can be corrected before the package is shipped. Of course, whatever the issue, it is still hoped that by addressing the first challenge above, recalls will be issued quickly, as records are also easily retrievable.

Challenge 4: Machining parts fast enough to keep production running is also a challenge, especially given growing sales and the need for a higher degree of customization.

Solution considerations: No matter how many different car models (and how many different combinations of parts) are required to complete a custom order, having a camera and scanner running constantly is the only way to ensure successful decoding many times per second.


If current machine vision cameras and fixed-mount industrial scanners are not meeting production needs, or are difficult to use and manage, then it’s time for a change. In an effort to fill these capability gaps, enhance process control, and simplify and speed up everything from quality control checks to non-compliance investigations, automakers can build on existing systems by integrating new cameras or scanners. Make adjustments, or better yet, invest in a new industrial automation system. Many customers are amazed that Zebra's dual-mode devices can be applied in many different ways through a single software platform, while learning that as operational needs change, it is possible to scale the solution or move between vision and barcode scanning applications Get peace of mind while switching easily.

Leave a comment

Your cart

×