Image Capture Central
Industrial cameras are the core component of many machine-vision and image-processing systems. Our annual guide to finding the right camera for an application—Vision Systems Design’s Worldwide Industrial Camera Directory—begins after page 20 of this issue. Cameras from more than 160 manufacturers and dozens of distributors in North America, Europe, and Asia are listed, identified by manufacturer, product name, sensor type, scan type, resolution, spectrum, interface, and data rate. This Directory will be widely distributed at trade shows and available on our web site atwww.vision-systems.com.
Selecting the best type of camera for a specific machine-vision system requires more than understanding camera specifications, however. Other important factors that must be considered include the proper choice of lighting, optics, and software. For this reason, the technical articles included in this year’s Camera Directory show how lighting can be used in a number of different ways to enhance image contrast, how telecentric lenses improve the accuracy of parts inspection, and how to test and calibrate imaging systems.
Further directions
Once machine-vision cameras capture images and transfer them to a host computer, they must be processed. But what sort of image processing is really required? In this issue, we challenge some of the most important assumptions about both the nature of such image-processing algorithms and the tasks now faced by software developers.
In the first article, Ganesh Devaraj and his colleagues at Soliton Technologies examine the image-processing algorithms used in smart cameras. The article examines how frequently typical algorithms are used and suggests that a majority of smart-camera applications can be solved with only a small number of algorithms that can be learned quickly and used effectively.
Taking up this challenge, editor Andy Wilson explores the question of benchmarking the performance of machine-vision software packages and tasks vendors to develop a single, albeit large, image test file. Based on such a test, systems developers could rapidly analyze the speed of many—if not all—of the most important algorithms used in today’s machine-vision applications.
Although you will find this Camera Directory to be an invaluable resource, our staff has recently added even more information to our web site. For example, our video library now contains updated tutorials, including several that were recently presented during NIWeek on advanced optics for machine vision, multicore vision, and trends toward parallel processing, and advanced imaging preprocessing.
Although no single article, presentation, or directory can answer all the questions related to designing a machine-vision system, we hope that a combination of such resources will provide a more complete understanding of the tradeoffs involved.
W. Conard Holton, Editor in Chief
[email protected]