Starting point for vision success
By Larry Maloney, Contributing Technical Editor -- Test & Measurement World, 11/1/2008 2:00:00 AM
|
|
|
If you’re considering machine vision, it’s tempting to get caught up in camera choices, lighting schemes, and setup programs. But you should ignore such thoughts until you have thoroughly analyzed your application, according to David Dechow, president of Aptúra Machine Vision Solutions, a Michigan integrator.
“Many vision applications fail or have only limited success because people neglected this important analysis stage,” said Dechow, who has spent 25 years as an engineer, programmer, and manager in the vision field.
Dechow explained that even before this analysis begins, you need to understand what your process needs with respect to inspection, and you should also understand your company’s quality requirements. How much is your company willing to spend for improved inspection? And who will be the project’s champion—or will decisions be left to a committee, a far less appealing approach?
For application analysis, Dechow uses questionnaires with his clients to probe a series of issues surrounding the current production process, inspection criteria, and part descriptions. Among masny key questions: What in the production process is causing the errors? How are bad parts rejected under the current system, and what remedies have you already put in place to improve production? What features of the part need to be inspected? Do you need to read a bar code?
 A camera checks surface quality on electronic components. You should postpone selecting components for vision systems until you’ve thoroughly analyzed your application, according to integrator David Dechow. Courtesy of David Dechow. |
Dechow also is a big believer in having a thorough description of the part to be inspected. This description should involve physical details about geometric structure and features, as well as all possible part variations, including color, size, materials, and surface finish. And will the part change over time?
“The part description, including drawings and photos, is essential because your inspection system must be flexible enough to handle all possible variations,” explained Dechow.
Equally important is a thorough understanding of the material-handling system that delivers the part to the inspection station. What will be the speed and throughput of the system—and the positional variations of the part? Will there be shock and vibration? Machine vision is extremely sensitive to how a part is presented, and solving that challenge can be even more difficult than designing the inspection system. Said Dechow: “The part needs to be positioned properly for the camera, and accomplishing that in a production setting is a lot harder than doing it on a flat bed in the lab.”
Other key questions in application analysis involve the operator interface. Do you want the machine operator to see results of the inspection or have access to the machine-vision control system? Does the vision system need to interface with a PLC (programmable logic controller) or a robot?
Finally, the automation team must agree on what will constitute success in a new inspection system. That entails defining specific acceptance criteria for a part, a process that typically includes defining all the attributes of a good part.
It’s only after you’ve taken these and other application analysis steps, said Dechow, that you’re ready to move on to the project specification stage, where you actually design the inspection system and choose the hardware and software you’ll need.
- 0 rated items found.
Featured Company
Most Recent Resources
- Load, Pressure and Torque: Sensor Selection, Technology and...
- Mass Flow Versus Volume Flow
- Aries’ New Fine Pitch Bump Adapters Adapt BGA Devices of Any Pitch...
- New BGA Switch-A-Pitch Adapters from Aries Now Adapt 0.50 mm Devices...
- The DataFlex-1000A Flight-Capable High-Speed, High-Capacity Data...
View All Webcasts
