Q: How would you assess the progress in getting companies to adopt machine-vision solutions?

A: Frankly, this is difficult for us to answer. Our customers are the ones convincing end users to adopt machine-vision. If you take a look at the financial results of the major component players, you’ll see that sales are not growing by leaps and bounds. But sales are progressing slowly, and volumes are up. In addition, there is no shortage of new opportunities. The semiconductor-inspection market is still lucrative, while other markets like robotics are coming on strong.

Q: What’s behind the growth in robotics applications?

A: The increased interest in vision-guided robots can be attributed to “flexible manufacturing,” where an assembly line must be able to make different products. A robot with vision can “react” to a given situation and instantly accommodate the needs of a particular product run. Smart cameras nicely complement robotics applications because the camera can control the robot itself, very much like a remote PC.

Q: What are the chief barriers that slow the adoption of machine vision?

A: Price still plays a significant role, but that is changing as the cost of technology decreases. Historically though, there are users who have experienced significant difficulties at the deployment phase, and that led to a lot of poor-quality systems, which in turn damaged the overall perception of vision systems. What that means for machine-vision vendors is that there are some users who need to be convinced that machine-vision solutions can be set up and maintained easily.

Of course, the industry has matured in leaps and bounds, so the technology has improved. Vision libraries, such as the Matrox Imaging Library, are also including more higher-level tools to remove some of the complexity of implementing machine-vision systems.

Q: How can vision companies make systems more attractive to customers?

A: Vision products must be robust and easy to use, and it’s the software that’s the key. You can’t take advantage of the latest computer chipsets or frame grabbers or the highest data-rate 2-D and 3-D cameras without appropriate software support.

That said, software tools are evolving, too. Early applications relied on low-level processing primitives such as filters and transforms to extract the required information out of an image. Today, those primitives have been built into high-level functions. We now have ready-made tools that help developers build more robust applications in far less time and at a lower cost.

Q: How is Matrox addressing ease-of-use concerns?

A: We’ve addressed the ease-of-use issues largely with our Matrox Design Assistant interface, which is designed for use with the Matrox Iris E-Series smart-camera family. Design Assistant is an intuitive flow-chart-based development environment that helps developers configure and deploy machine-vision applications without programming. Its development environment provides access to a comprehensive set of highly efficient and field-proven image-analysis and measurement tools.

Q: How are you helping your customers adapt to the variety of machine-vision protocols?

A: The Matrox Imaging Library has been platform-independent since its first release in 1993. One of our very earliest goals was to support different interface standards, whether through our hardware or via third-party products for both the GigE Vision and 1394 interfaces.

Read the continuation of this interview.