Winning machine vision's "Red-Queen's Race"

Inspection and test technologies, by their nature, always lag one step behind the products they are trying to inspect or test. Even an understanding of guiding principles like Moore's Law (circuit density doubles every 18 months) does not permit an organization to always anticipate the pace of progress. We find ourselves in a perpetual "Red-Queen's Race" (from Alice in Wonderland)—running as fast as we can just to stand still.

Nello Zuech, president of Vision Systems International (Yardley, PA; www.imagelabs.com/vsi), points out that the light at the end of the tunnel is not, in fact, an oncoming train. Zuech, who has written extensively on machine-vision technology and applications for the Automated Imaging Association (AIA; www.machinevisiononline.org), identifies the "good news" that the electronic industry is always in a state of flux (no pun intended). New board-level designs must accommodate ICs of ever-higher circuit density and ever-declining package size by increasing interconnect density. The "bad news" is that vision companies must constantly adapt their systems to handle the changes.

Fortunately, technologies at the core of machine-vision systems are meeting this challenge. Zuech says that even cameras with resolutions in the 1-Mpixel range have become passé. At The Vision Show East (May 4–6, Boston, MA), sponsored by the AIA, vendors exhibited cameras capable of resolution at 4, 6, and even 11+ Mpixels.

These cameras featured CMOS and CCD imagers, a choice of frame rates, digital zoom, and connectivity via Camera Link, FireWire, USB 2.0, and Ethernet. Prices for many were low enough that they could be used in cost-effective machine-vision systems.

In addition to developments in camera technology, vendors at the show displayed LED illumination arrangements in virtually every format. Compared to conventional lighting, LEDs help to reduce total machine-vision-system cost of ownership by reducing both the maintenance associated with lamp replacements and the cost of the lamps themselves.

Processing the enormous quantities of raw data generated by high-end cameras has often been a challenge. Fortunately, according to Zuech, the speed and memory capacity of the latest computer technology allows new compute-intensive algorithms to handle the increased load.

New vision systems can perform rigorous, on-the-fly decision making without compromising production throughput. In addition, Zuech contends that systems based on today's computer tools are easier to train for board inspection and that they generate fewer of the false calls that have traditionally plagued machine-vision systems used in electronics manufacturing.

The need to adopt machine-vision tools for electronics inspection is undeniable. News from The Vision Show East suggests that many products are now up to the task.