Ordering from the vision menu
By Larry Maloney, Contributing Technical Editor -- Test & Measurement World, 5/1/2009 2:00:00 AM
|
|
|
With capital investment funds at a premium, companies can ill afford to make a mistake on machine vision. “Very often when I get called in, companies are fairly certain about what they want to do, but I have to talk them out of it because the approach just isn’t right for their application,” said Perry West, founder and president of Automated Vision Systems, San Jose, CA.
With 30 years of experience in the field, West always asks companies to first state in a single sentence what will constitute success in the vision system they plan to install. For some, it might be labor savings or scrap reduction; for others, it could be an increase in inspection throughput. Only with a goal in mind, said West, can you start to define the actual specifications of a system.
|
West also has devised a simple taxonomy (see figure) that helps companies get closer to their ideal system. “In broad terms, R&D people gravitate to general-purpose vision systems because of the flexibility,” explained West. “End users, such as a semiconductor fab, often look for application-specific systems that have proved successful throughout their industry.”
What’s the best option among general-purpose solutions? West offered this perspective on the three primary choices:
-
Component. While it may offer cost savings and greater flexibility, a component-based system can be difficult to implement because you must buy a PC, frame grabber, camera, software, and light source. Integrating these elements can baffle vision newcomers.
-
Packaged. To avoid the complexity and compatibility issues of component systems, some vendors integrate vision components into modules that include, at minimum, a camera, a processor, and an operator interface. The downside of these packages: higher costs.
-
Smart cameras. Because the image processor is typically built into the camera, users get an easier-to-implement system with a smaller design footprint. Overall, it’s a cheaper alternative to both the component and packaged approaches.
When general-purpose machine vision won’t do, West suggests three alternatives that deliver application-specific solutions:
-
Integrated. Here, a system integrator tailors a complete vision system, including programming, to a specific end-user application. The outside expert may also provide engineering for parts handling, robotics, and motion control. At first blush, this solution might appear expensive, but it can substantially reduce risks—and costs—in complicated applications.
-
Product. You gain cost advantages with this approach because you’re buying a catalog item, typically a plug-and-play vision module that can be dropped into a manufacturing line. These products often include part-handling or conveyor capability.
-
Embedded. A prime example would be an OEM designing a smart camera into a machine programmed for a specific task, such as inspecting leads for integrated-circuit packages.
What’s hot among these vision choices? “Clearly, it’s the smart camera,” said West. He also sees strong interest in turnkey vision systems that meet the needs of specific high-volume applications, such as semiconductor manufacturing.
|
No related content found.
- 0 rated items found.
Datasheets.com Electronic Parts & Inventory Search
185 million searchable parts
- Part Number
- Description
- Inventory
- Products
- Manufacturers
View All Webcasts
