Quality Magazine

Vision Systems Add Accuracy

February 25, 2010
Using machine vision enables manufacturers to inspect 100% of their products and ensure that defects don’t leave the production floor.

Vision systems align solar cells, check for cracks, edge chips, and detect material and process fluctuations that degrade solar cell yield and energy conversion. Color vision tools also can be used to sort cells by color prior to module assembly. Source: Cognex


Machine vision is a powerful tool for managing and improving product quality, and the cost of using vision, both in dollars and engineering time, continues to move in the consumer’s favor.

Today it is more affordable than ever to put machine vision to work on production lines where these systems can deliver powerful benefits. Consumers today expect perfect quality, even in low-cost items. Using machine vision enables manufacturers to inspect 100% of their products and ensure that no defects ever make it off the production floor.

In addition to eliminating defective parts, machine vision also helps manufacturers error proof the production process itself, so that fewer bad parts are made. Vision systems collect volumes of real-time process data that can help operators monitor, analyze and improve how the equipment is working. Using vision to alert floor operators about processes that are starting to move out of tolerance allows them to correct the process before defects appear.



A Growing Range of Options

Vision can be used in many ways to reduce defects and ensure quality. The most common is inspection, in which vision systems inspect each item and make sure a process was performed correctly, such as verifying that caps and labels are present on every bottle.

Vision also can add greater consistency and accuracy in assembly processes with precise robot guidance and part alignment, and by verifying that critical dimensions are within tolerance.

It can be used to identify items during production, either by physical appearance or by reading an identifying code marked on each item, to ensure the right part is being used in an assembly, or that the labels being applied to a container match the contents.

The vision industry now has 30 years of hands-on experience with manufacturing applications in a range of industries. This increased understanding of manufacturing processes has driven machine vision companies to build a diverse range of products suited to the specific quality control and inspection needs of manufacturers.

Manufacturers today have an unprecedented wealth of options when it comes to selecting a machine vision system or vision sensor; whether they need programmable software, a plug-and-play smart camera or specialized ID reader, there is a system available to do the job, and these systems can be purchased in a wide range of resolutions, speeds and price points to meet the application’s needs.

Today, even the most affordable solutions can be extremely powerful. There are dedicated presence/absence vision sensors that can inspect multiple part features simultaneously at speeds of up to 6,000 parts per minute. These inspection tools are the right option in many cases where a general purpose vision system with more capability is unnecessary.



In the food industry, vision systems need to be able to withstand moisture and solvents used during wash down. Rugged die-cast aluminum and stainless steel cases that can withstand the punishment of vibration, sealed M12 connectors and protective lens covers for IP67 and IP68 rated protection against dust and moisture are options that keep systems running in harsh factory environments. Source: Cognex

Factory Communications

The adoption of component object model (COM) technology and ActiveX controls starting about 10 years ago greatly simplified how customers linked together devices from many different companies to have them all communicate effectively.

More recently the transition to Ethernet and adoption of open network protocols like EthernetIP, Profinet, ModbusTCP and MC Protocol has allowed easier integration of sensor products into the control platform.

This improvement in factory communications delivers some significant benefits to operators. Vision systems collect data that factory floor managers can use to analyze their work in process or quality metrics. Today, it is possible to communicate that data easily to other factory controllers to help speed correction of process errors or archive quality information.

Imagine a vision system inspecting label placement on bottles. The vision system can be set to send position data back to a controller, which can be programmed to send an alert to operators when labels begin to move out of tolerance, effectively enabling operators to take corrective action before the defect occurs.

Also, it is now possible to easily save defect images and other graphics directly from vision systems to remote file servers. Stored images can be used for future failure analysis and troubleshooting.

The trend toward GigE will further benefit vision systems as image resolution increases and production line speeds increase. Bandwidth improvements also have enabled companies to transmit larger volume data at high speeds. Increasingly, quality engineers have an unprecedented wealth of qualitative data-not just pass/fail-to work with in order to analyze and improve their processes.

Brand reputation is more important than ever, not just with consumers. Large buying organizations often refuse to do business with any supplier whose quality has been questioned. Vision systems give manufacturers the tools they need to guarantee product quality and provide proof of product quality to their customers.

Traceability

As manufacturing becomes more complex, more companies are implementing traceability programs that enable them to provide a fully auditable trail of how each individual item was handled and processed. Image-based industrial ID readers are a reliable and cost-effective way to meet requirements for item-level traceability.

Image-based ID readers have the ability to read a range of codes, including both traditional 1-D bar codes as well as 2-D data matrix codes, which are increasingly the code of choice for traceability applications, particularly in the pharmaceutical and medical industries. As traceability standards are set in a variety of industries, vision companies are keeping pace with specialized interfaces for more easily capturing required data, adding other capabilities such as code verification, and providing a range of form factors, from fixed-mount to handheld units and price/performance points.

Networking

Network management tools now exist that make it easy to configure and manage multiple vision systems from a single control station.

Some operators can have as many as 100 vision systems or vision sensors operating on a dozen different lines. Managing all these sensors from a centralized control unit saves time and effort when systems need to be modified during changeovers, and it reduces the labor required to manage vision applications.

And, because information from the vision system can now be imported directly into almost any factory HMI or Scada control interface, it eliminates the clutter and maintenance of separate display terminals.

Better Color Tools

Color tools have caused operators some frustration in the past due to inconsistent performance. The good news is that today’s color tools have improved in both accuracy and ease of use. New 24-bit color tools give vision systems greater sensitivity to subtle shading differences and makes color tools more accurate and tolerant of variable lighting conditions.

The improvement in color capability enables many new and valuable inspections, since some products, particularly in the food and pharmaceutical industries, only can be distinguished based on color. New color tools also can be trained to find not just one color but a combination of colors and color markings.

Line Scan Image Acquisition

Vision system performance is greatly influenced by the type of camera used to acquire the image of the object being inspected. Higher resolution cameras enable better visualization of very small or subtle defects. However, not all parts to be inspected match the aspect ratio of traditional 2-D image sensors-think computer keyboards to dental tools.

One solution to this problem is to use a line scan camera. Smart cameras using line scan image acquisition can provide higher resolution at a lower cost than traditional high-resolution area cameras because an image sensor can be more easily matched to the dimensions of the part. For example, a line scan camera can easily acquire a 2K by 8K image of a long rectangular part. And because line scan cameras typically have very short exposure times, maybe 10 to 100 times faster than an equivalent area scan camera, they work without using strobe lighting, which can be annoying or disruptive to manufacturing line workers.

One of the benefits of line scan cameras is that they can be used where space is tight, such as in an existing piece of equipment or on a space-constrained line where it is not feasible to evenly illuminate an object or to even “see” the whole object with a traditional 2-D high resolution camera.

Line scan cameras also have other benefits, such as the ability to inspect cylindrical parts. Vision systems offer improved product quality as well as proof of product quality to manufacturers' customers.