Machine vision suppliers said that North American revenues increased 25.5 percent between 1999 and 2000, surpassing the expectations of the Automated Imaging Association for 2000 and almost those for 2001. The association's 2000 machine vision market study predicted 13.8 percent annual growth through 2004, but significant growth in semiconductor and electronics capital spending last year resulted in revenue increases of almost twice the predicted levels. North American revenues for machine vision components exceeded $1.57 billion in 2000, with value-added revenues raising that figure to $2.11 billion, according to the imaging group's 2001 market survey. The survey estimates world machine vision revenues at $6.3 billion, with Japan at $2 billion, Europe at $1.5 billion and the rest of the world at $700 million. Nello Zuech, president of Vision Systems International in Yardley, Pa., revealed the preliminary survey results just hours after a more sobering talk that suggested the 2002 results may not be quite as rosy. Slow start in 2001 Pat Lamey, senior director of strategic directives at Applied Materials Inc. in Santa Clara, Calif., led off the association's annual business meeting with news that this year appears to be about 7 percent below expectations so far. Lamey said inventory buildups in telecom and slow personal computer sales have made electronics and semiconductor customers re-evaluate their spending, including for inspection equipment "2001 doesn't look too good," he said. With semiconductor and electronics applications still the key drivers for machine vision, it took some optimism for Zuech to predict 14.1 percent annual revenue growth for the next five years and 14.8 percent growth in the number of units. Zuech noted that 2000 unit sales rose particularly sharply -- 32 percent -- in applications other than semiconductor and electronics, while the corresponding revenues were flat. He said the survey indicates that new, less-expensive vision systems -- smart cameras and embedded systems -- are creating opportunities in many markets. A quality engineer might be reluctant to recommend spending tens of thousands of dollars for something untried, but a $2000 system is "not a career-ending decision," Zuech said.