I am a camera
If there is one central theme that has been pervasive in the development of image-processing and machine-vision systems it has been the camera itself.
If there is one central theme that has been pervasive in the development of image-processing and machine-vision systems it has been the camera itself. In the more than 30 years since the first CCD camera was developed, many significant developments have occurred. Among them is fabrication technology, which has allowed companies such as Kodak to announce 35-mm optical-format, 16-Mpixel, 3-frames/s CCDs targeted at industrial, scientific, aerial, and security markets.
More radically, the introduction of low-cost CMOS imagers has generated a host of consumer and niche products for automotive and high-speed applications. Cameras based on unipolar and bipolar technologies that digitize visible, infrared, and ultraviolet sensors are now being used to digitize images and glean data for controlling automated machines and industrial processes.
As this issue shows, such cameras have been put to a multitude of uses, ranging from low-cost “smart” sensors that enable simple tasks to high-end cameras with frame grabbers that drive complex automation equipment. In our Product Focus article, editor Andy Wilson shows how established vendors of industrial-automation equipment are providing system integrators with smart sensors to help ensure that specific quality processes have been “error proofed.”
More complex vision systems and processes are described in articles on the manufacture of forestry and automotive products. Lutz Kreutzer of MVTec Software describes a system developed for an Italian lumber mill to inspect milling, gluing, and compression of laminated beams. And Andy Wilson writes about an automated, modular catalytic-converter inspection system that relies on a mix of linescan and area-array cameras.
LIFE IS A PUZZLE
The camera is only one element in a system for solving industrial challenges-or academic ones that may be applied to industry. As our cover story relates, the challenge for a group from Carinthia University of Applied Sciences was to build an autonomous robotic system that detects the state of a Rubik’s cube and then computes and controls how it is solved. A standard FireWire camera acquired the image, but it was the sophisticated image processing and analysis, followed by the operational computation, that enabled the robot-control system to unscramble the cube.
Off-the-shelf machine-vision components were used in all the applications described in this issue. But they were often customized to address each system challenge. As Paul Bourget from LumenFlow, the subject of our interview this month, says, standardized products that can be readily customized to create flexible, integrated systems are the key to solving future puzzles for machine-vision and image-processing applications.
W. Conard Holton
Editor in Chief