![](/QUAL/Home/Images/qu0207-FT2-test-p1.jpg)
As integral as machine vision has become to such industries as food and beverage packaging, printing and automotive parts, industry analysts projected its across-the-board growth in the manufacturing sector through 2006 at about 9% annually. What have remaining manufacturers been waiting to see in order to consider it an indispensable asset? Reports published over the past three years consistently point to:
• Uncomplicated, easy-to-use technology requiring no on-site expertise
• Systems that measurably contribute to consistent quality improvement, lower scrap rates and enhanced operating efficiencies upstream and down
• Integration within the manufacturing process to supply information that improves process control
• Industry- or application- specific systems that inspect for finer, obscured and more specialized defects-for example, 360-degree inspection and multiple inspections on higher-speed lines
![](/QUAL/Home/Images/qu0207-FT2-test-p2.jpg)
The stakes are high. Manufacturers who achieve dramatic, permanent reductions in defect rates can save millions of dollars per year. On the contrary, a single defect unleashed in the marketplace can have the exact reverse effect-or worse. Add to that the extensive process problems one out-of-tolerance manufacturing component can cause when not immediately detected and corrected. Those are the issues that today’s machine vision systems are addressing now.
An Industry Example
Consider today’s beverage producers, for instance. While inspecting containers for structural anomalies, damage and debris is essential for can and bottle manufacturers, it is a sound business practice for beverage producers as well. Problems that may have occurred in transit or storage, or those missed in the manufacturing process, can become issues for the filler. Examples include contaminants or objects that can enter containers, glass bottle chips or cracks and can dents.
![](/QUAL/Home/Images/qu0207-FT2-test-p3.jpg)
To prevent process and equipment downtime caused by container problems-and even far more costly consumer concerns-many producers today still rely on human inspection and sampling methods.
However, today’s machine vision inspection systems are performing 100% inspections at line speeds of up to 3,000 containers per minute and automatically rejecting defects from the line. Importantly, they also can detect many smaller and even hidden defects that no other method can.
The avant-garde of beverage containers, such as the latest aluminum bottles, brings a unique set of challenges to quality assurance. The majority of aluminum bottles now are crown-cap type. The formation of their neck and curl, or cap-seal, areas is a complex process, involving up to 17 manufacturing stages. The biggest potential problems can occur in those areas, particularly splits or deformities in the curl.
The underside of the curl, a likely place for a defect, is a challenge to inspect. The possible consumer injury and compromised seal integrity that could result are concerns to both the filler and container manufacturer.
Other main concerns are damaged or malformed containers that cause line and equipment problems during filling, and unacceptable neck decoration defects.
For the leading manufacturer of these containers, running up to seven lines even at relatively slow speeds of 150 to 200 containers per minute, human inspection was not a viable consideration. An inspection system with a digital four-camera configuration for 360-degree inspection of the curl area, together with custom-designed LED lighting, optics and intuitive operator interface, provides an effective and manageable solution that meets the manufacturer’s needs and critical standards.
Rapidly developing in aluminum bottles are recloseable screw-cap designs to accommodate waters, wines, juices and sodas. These require inspection of tight thread areas on the neck and their narrow neck interiors. There also is the trend to make aluminum bottle walls thinner to conserve costs of material, handling and shipping. That will present additional concerns in maintaining consistent packaging integrity and inspecting for it, a current focus of leading machine vision suppliers.
![](/QUAL/Home/Images/qu0207-FT2-test-p4.jpg)
Resolving Concerns
Today’s remaining reasons for indecision about machine vision are being addressed. The best news first-the technology is continually getting faster, smarter and less costly.Machine vision consists of lighting, optics, video cameras, and computer software and hardware, including operator interface. All of which, like electronics everywhere, are benefiting from faster processors, more sophisticated software and hardware components that are less expensive to make yet markedly more robust, durable and reliable.
The optimal current technology includes dual-processing CPUs that allow multiple cameras to be run from one processor, reducing the cost per camera; digital gigabyte Ethernet cameras with longer transmission links that can be positioned farther away from the CPU, thus increasing the number of lines that can be inspected with one CPU; fully software-programmable smart cameras containing their own digital processor chips to handle more of the data processing and provide greater inspection flexibility, convenience and economy; and highly efficient, better-performing hardware technologies such as LED lighting and higher frame-rate camera capabilities.
But that technology is only better when it is incorporated with concurrent advantages available, including:
• Application-specific system development. Any task goes better when operators use the tool designed for the particular job. For reliable 100% inspection of one’s unique areas of interest, at certain line speeds, in those plant configurations, and to best serve a company’s processes, the most effective machine vision is application-specific. In being the most effective it should prove to save companies the most and, overall, cost the least.
• Up-front expertise and intuitive user interface. The system’s development should be expertly provided right up through complete setup. A machine vision inspection system must be quick, simple and mistake-proof to use. The leading systems have colorful graphic touch-screen controls that let operators key-in inspection parameters in five minutes or less, make sensitivity changes on the fly, plus have the system check the acceptability of the proposed changes, minimizing uncertainty and risk. The system should facilitate all of this without interrupting the current inspection process.
• Integrated process-control support. Earning more than their original keep, today’s in-line inspection systems are monitoring, counting, reporting and signaling problems. A new advanced process monitoring technology is going beyond conventional capabilities such as process feedback and statistical analysis. When inspection systems with this intelligent-machine software detect a defect trend, they also are able to recognize the exact upstream manufacturing component causing it, and can immediately alert or page personnel, send notification to a device, display it on a marquee or shut down the line, according to plant’s requirements.
More than a method or a device, machine vision inspection is technology that is becoming an integral part of the manufacturing process. This goes beyond supplier-industry responsiveness. The machine vision industry has been watching and, to a great extent, experiencing the challenges its customers face in adopting ever-higher quality standards and meeting unprecedented performance levels. And it knows that process is continuous.
Whether machine vision inspection is well in place for the operation or a company has tabled exploring it for a time, the current environment, recent developments and evolution of the industry warrant a look into how machine vision inspection could be benefiting operations now. Q
Tech Tips
• Machine vision consists of lighting, optics, video cameras, and computer software and hardware, including user interface.• Today’s machine vision inspection systems are performing 100% inspections at line speeds of up to 3,000 containers per minute and automatically rejecting defects from the line.
• Machine vision systems can detect many smaller and even hidden significant defects that no other method can.