Anticipating the learning curve
Andy Wilson Editor at Large
Today, systems integrators are confronted with a myriad of technologies with which to implement image-processing systems. These technologies include, for example, digital cameras, PCI-based frame grabbers capable of transferring images at video rates to host PC memory, DSP-based image processors, Windows-based development tools, erasable optical memory, and low-cost color monitors. Faced with this array of technologies, system integrators must deal with bewildering choices. What`s more, with the decreasing time to market for OEM products, they must also decide on which products to use in designing a future-generation imaging system based on late-breaking availability.
In structuring imaging-system designs, an increasingly important requirement is the anticipation of users` future needs. Consequently, system designers must try--within reason--to incorporate new products into systems under development. Primarily, this expectation has been fostered by the rapid development of PC products. Indeed, most PC vendors are already evaluating Microsoft Windows 98 (albeit still in beta test) and the latest version of Windows NT for incorporation into this year`s computer systems.
For integrators of image-processing systems, the choices become complex because of the variety of available hardware and software products. But, just as with the choice of operating systems, system integrators must carefully consider that OEM suppliers are furnishing current and leading-edge product technologies. Simply choosing a monochrome camera, digital interface, PCI-based frame grabber, and software from several OEM suppliers could solve a particular machine-vision problem. But do these OEM suppliers have a road map of new product technology introductions planned that systems integrators can count on for future system improvements? And, will the software required to develop such an image-processing system be reconfigurable when new hardware and operating systems become available?
Certainly, these important questions are being considered by OEMs. But, perhaps the most important question is the value-added product perception by customers and end users. For example, in the case of PC technology, most PCs now look alike. They generally incorporate Pentium II processors running at various speeds with several megabytes of memory, gigabyte drives, and CD-ROMs. The only discriminating factors appear to be price and bundled software. Moreover, the value added by resellers of PCs has fallen dramatically as the ease of configuring systems has increased.
Fortunately, developing high-performance image-processing systems still requires considerable engineering expertise and experience in choosing the proper OEM components. Indeed, although the hardware and software-development tools for producing face-recognition systems, for instance, might overlap somewhat with those involved in web-inspection systems, the imaging expertise required to develop vision systems remains as in-depth and diverse as the markets themselves.
But, just as PC components are being built to plug-and-play standards, OEM imaging components are following suit. Although this standardized approach will lower the cost of configuring imaging systems and reduce end-user costs, systems integrators will have to add more value to meet customers` needs. Only by keeping abreast of product technology changes in every aspect of OEM development will systems integrators be able to deliver what their customers need now and in the future.
