2001 Should End Well for Machine Vision Suppliers

With the economy slowing and some sectors reducing capital expenditures, it's going to be a scary ride for the first half of 2001 for machine vision vendors. But, in the end, demand for zero defect production in the manufacturing sector will prevail, resulting in machine vision system sales equal to or slightly better than 2000.

'Sales in 2001 will be better than 2000 with the disclaimer that we're all hoping that the slight economic downturn [at the start of the year] doesn't turn into a huge economic downturn,' says David Dechow, president, ISRA Vision Systems/Insight Integration, Lansing, MI.

'In our experience, particularly going back to the last downturn in 1989-1991, capital purchases for automation equipment tended to keep pace,' he notes. 'That's because heavy industry tends to try to gear up with capital purchases during times of slow down,' he explains.

'If the economy tightens up, we think our technology and industry should have a bumper year because we can reduce costs and displace headcount to more valuable tasks than sitting on-line inspecting,' agrees Robert Steinke, president of DVT Corp., Norcross, GA.

The scarcity of labor, pressure on pricing and the need to eliminate as much waste as possible also encourages adoption of machine vision since it is an essential tool when human inspectors are in short supply and manufacturing costs need to be minimized without compromising quality.

'Competitive pressures today are very intense,' says Fred Molinari, president, Data Translation, Inc., Marlboro, MA. As a result, 'machine vision will be more sought after in a lot of industries, which really haven't been big users up until now,' he says.

Putting pressure on results will be a softness in semiconductors and electronics, the major user of machine vision systems. 'Clearly, the semiconductor and electronics industries are very influential,' says Justin Testa, senior vice president of marketing at Cognex Corp., Natick, MA.

Since more than half of the machine vision systems sold are used in these two sectors, 'it's where the greatest opportunity exists,' agrees Joe Christenson, president of PPT Vision, Inc., Eden Prairie, MN.

'The question right now,' says Testa, 'is, 'How deep of a slowdown will this sector experience?'' However, since electronics and semiconductors are in much broader usage than in past downturns, 'perhaps this slowdown won't be as steep or as lengthy,' he notes, hopefully. 

Moving beyond electronics and semiconductors, 'We still see very strong demand for automation technologies in a number of key industries,' says Testa. 'The pain of manufacturing is so high, they have to automate,' he concludes. Particularly strong interest is being seen in a number of sectors including automotive, fiberoptics, food/beverage, transportation, document processing, security, general manufacturing, packaging, plastics, printing and converting and pharmaceutical/medical device/healthcare. In fact, packaging, and pharmaceutical are so strong, 2001 sales in these areas could overtake sales to the automotive industry, traditionally a larger market.

'The application area we see as having the greatest growth potential is the whole realm of security applications,' says Philip Colet, vice president of marketing at Coreco Imaging, St. Laurent (Montreal), QE. This includes biometrics, face recognition, people tracking and traffic monitoring. By making it possible to reroute drivers around problem areas, the latter has 'huge implications for energy savings and pollution control,' says Colet.

Fiberoptics is such an active area that Cognex has set up a team of engineers focusing on this sector. Since this is the medium that allows the transmission of data at very high rates and more and more devices are designed to be web-enabled, demand has skyrocketed, catching manufacturers somewhat off-guard. The rising growth rate makes it difficult for manufacturers, who have been using a largely manual assembly process, to keep up and is forcing them to automate, quickly.

Another emerging area is part traceability, which relies on direct marking on parts via etching, stamping or printing, generally with a 2D matrix code. Depending on the part and marking method, the contrast of the code to the background can be very low and presents a challenge to reliable and repeatable reading. Just starting to arrive on the market are self-contained, Ethernet-equipped readers capable of examining up to 30 codes per second and transmitting data to a host system.

Another application with huge potential is healthcare where machine vision can enhance diagnostic images, track biometric data, and provide guidance for robot and laser surgery.

Finally, the rapid development in technology over the past seven years or so means a growing number of sales are replacements of existing systems. Upgrading is done for the same reasons as an initial purchase: reduce cost, improve quality, boost throughput. 'Speed is a huge driving factor,' says Dechow. Higher speeds and more robust algorithms allow today's systems to inspect faster, perform more tasks and/or accomplish functions not previously possible.

Increasingly, machine vision systems are relying on digital technology, particularly digital cameras and interfaces. Digital signal processors, for example, boost processing power at the CPU level and minimize the need for special hardware. Higher processing power enables development of more advanced vision algorithms to perform more complex tasks like 3D measuring and pattern defect inspection and classification.

'More processing power means vision programmers and engineers can create even better software to do more automated things, while making systems more reliable and easier to set up. In short, higher performance, lower cost processing,' says Testa.

With Ethernet rapidly becoming the primary networking architecture for the factory floor, there is wider use of it to network machine vision systems to other devices. Another popular feature, web enablement, provides remote access for programming, monitoring and maintenance.

With today's higher processing power, faster speeds, better software tools and simpler user interfaces, machine vision systems are easier to install and operate and more flexible. These developments are moving the technology into new applications and multiplying its use in existing niches. 'Customers are looking for ways not to just inspect quality, but to use machine vision in an environment of flexible automation,' says Dechow.

To spur adoption, suppliers are beginning to introduce vertical tools, especially designed to perform certain tasks, but flexible enough to allow some custom configuration. Integrated systems also are popular since users don't have to struggle to make a group of components work together to perform the vision task. 'Ease of use is critical,' says Fred Molinari, president, Data Translation, Inc., Marlboro, MA.

By combining more functions in a component, smart cameras and CMOS technology are helping make integrated systems a reality. CMOS sensor-based technology in cameras not only simplifies operation, but also tends to lower costs and improve performance.

Lower costs, greater ease of use and network compatibility allow placement in many more places on factory line. 'The product that combines performance with low cost and ease of use is really a key enabler for manufacturers to implement machine vision automation in their process,' says Testa.

Wish list

'The biggest barrier to more widely proliferated use of machine vision technology is in illumination and optics,' says Christenson. Improvements in these areas simplify installation, configuration and operation and encourage wider usage of machine vision.

DVT's Steinke would like to see software capable of providing depth perception, peripheral vision and night vision. While some of these capabilities are available now, the software tends to run too slow to be practical.

Another desirable advancement, neural net technology, would bring machine vision closer in capability to human sight. Although it exists, it's not widely adopted and still needs to be combined with scaler technology to perform fine measurements.

Dechow predicts strong growth for color systems, which will continue to decline in cost. 'Ultimately, once the price is in line, there's no reason to do anything in black and white,' he says. 'If price points reach the right level, applications for robot guidance and part identification are going to grow because the customer base, particularly in heavy industry, wants more support in the area of flexible automation.' he concludes.