Advanced Imaging


Advanced Imaging Magazine

Updated: January 12th, 2011 09:49 AM CDT

Real-World Reliability

Designing a reliable inspection system with tools that can predict and model system variations
National Instruments
Permutations of an inspection image that depict possible system variations from camera position, changes in lighting and blur which can be analyzed using NI Vision Builder for Automated Inspection 2009.
NI Vision Builder for Automated Inspection 2009 offers utilities to model system variations.

By Kamalina Srikant
National Instruments

Intense global competition is putting pressure on machine builders and integrators to deliver systems with higher throughput, reduced operating cost and more reliability. With many of todayís industries continuing to automate processes, machine vision is showing up in a larger variety of applications to ensure product quality at faster production rates while reducing scrap.

Especially in the current economic climate, industry professionals want to know that when they invest in a system their money will go a long way, and in turn, work for them. With looming upfront costs associated with machinery, components and labor for implementing inspection systems, machine builders need to do their best to instill confidence in their end-users that the system they design will operate as desired.

For machine builders and integrators, the cost and time associated with validating systems as they are deployed for delivery can be very high. Because of this risk, they too want to be able to deliver systems that have undergone as much validation as possible in the earlier stages. This can go well beyond just having chosen the best lighting, lenses and cameras for the customerís budget and application, and only so much can be done to constrain the system and minimize the environmental variations.

In the real world, reliability cannot be quantified with an absolute yes or no rating. When designing an inspection system, a lot of feasibility testing often is required, and one of the biggest challenges in doing this is finding a way to properly reproduce the exact conditions in which the inspection will take place. Often, the difficulty is in foreseeing all of the conditions that can arise when a system is deployed. Decaying lighting, for example, can affect the captured images, thus affecting the results of image processing algorithms, even when the most well performing algorithms are used.

With the lack of predictive capabilities in a design, it can be difficult to reproduce the real-world conditions that will affect the inspection and tweak the image processing accurately.

1 2 3 4 next

Subscribe to our RSS Feeds