Printer Friendly

Machine vision to serve flexible systems.

"I see mounting interest in the use of machine visision for monitoring machine tools such as lathes, machining centers, and punch presses. Certainly machine vision will play a critical role in flexible machining systems--for verifying the presence of the part, measuring its location on the pallet relative to the programmed location, and then signaling any necessary changes in the numerical control program."

So said Perry West, one of the kickoff tutorial speakers at the third annual Applied Machine Vision conference held recently in Schaumburg, IL, a western suburb of Chicago. West is president of Automated Vision Systems, a consulting company headquartered in Campbell, CA.

Sponsored by Robotics International of the Society of Manufacturing Engineers (RI/SME), the conference drew 429 attendees, more than double the 192 who attended last year's conference in Memphis, TN.

The three-and-a-half-day conference featured illustrated talks by users and vendors on basic vision technology, a wide variety of applications, and prospects for the vision industry.

For the first time in the short history of the conference, this one was

accompanied by a miniatue trade show. Some 20 vendors displayed and demonstrated their hardware and software, and passed out literature. Interest evidently ran high, as the small exhibit room was packed throughout the two-hour show. Coolant critical

Expanding on this forecast, West noted that in machine-tool vision applications, the coolant presents challenges to system designers.

"Coolants or cutting fluids refract certain wavelengths of the visible spectrum more than others," he said. "This phenomenon, along with the attenuation of light by the liquid, may change the apparent size of the subject.

"As a result, the choice of coolant and control of its flow will become important elements in overall system design."

Fred Sachs, product manager for camera systems at General Electric Co, Liverpool, NY, noted that image density in solid-state camera systems has been rising rapidly.

"In 1973, we were working with arrays of 32 X 32 pixels," he said. "The count rose to 244 X 248 in 1978, 290 X 416 in 1983, and 512 X 512 this year. Some time next year, we expect to be working with arrays of 1024 X 1024 pixels.

"Everybody says they want more resolution," Sachs added. "We know how to build sensors with higher resolution; the problem is, they pick up more data than we know how to process."

He also predicted that the advent of color in machine vision is inevitable. "Color will greatly expand the market for vision," he said. Pertinent case histories

Several of the vision case histories described and pictured during the conference pertained to metalworking and metal-products manufacturing. For instance, Gary Ruzinski, applications engineer with Automatix Inc, Billerica, MA, described an installation at a company in Rochester, NY, where 14 cameras in two systems inspect engine connecting rods. The vision systems inspect for 14 different features in each of 15 parts-families.

Throughput is 1500 rods/hr, at 1.5 to 2.9 sec for each inspection. "If a camera spots an out-of-tolerance fault, the programmable logic controller is signaled to do no further work on that particular part," explained Ruzinski. "The vision system's controller can also do statistical analysis and plot trends in faults."

Joseph Otto, also an applications engineer with Automatix, described an unusual use of vision in robotic arc welding of reusable wear plates for railroad coal cars. Rather than tracking the seam for the welding torch, the vision system in this case simply measures the location of the part in the fixture relative to the programmed location. Deviations are then fed to the robot's controller, which adjusts the welding program accordingly.

Tolerances in the clamps may cause the parts to be mislocated by as much as 1/2" or [plus-or-minus] deg. The two-camera, three-dimensional vision system can identify, measure, and transmit deviations in 6 sec. Torch depth is determined to within 1.00 mm.

The user, Norfolk & Western, anticipates payback in one and one-half years or less. Lab developments

Two of the speakers gave details on results of research and development in vision application. Dr Fred Y F Cheu, a researcher in Manufacturing Engineering and Development at the General Motors Technical Center, Warren, MI, described the use of vision in automatic magnetic-particle inspection of connecting rods.

In the new process, the magnetic particles are coated with a resin containing a fluorescent dye. Two black lamps project ultraviolet light at low, raking angles. Equipped with filters that block visible and infrared rays, the cameras see only the ultraviolet reflections.

The production version of the inspection system will use solid-state sensing arrays and strobe tubes, Dr Cheu reported. Strobes put out more light energy than do conventional tubes or bulbs, and the use of strobes eliminates blurs caused by vibration or movement of the parts.

Dr Cheu added that the new technique offers advantages in determining the depth of flaws as well as their length and width, and that system accuracy is now at [plus-or-minus]2 percent false rejects.

Edmund Bangs, chief welding engineer at the Illinois Institute of Technology's Research Institute (IITRI), Chicago, IL, reported on work he has been doing in vision-aided infrared analysis for adaptive control of fusion welding.

Funded in part by the US Dept of Defense and several private companies, the project involves generation and analysis of thermograms that show the shape and areas of heat concentration in the developing weld puddle. "Users of TIG welding will likely be the first to enjoy the benefits of the new technique," said Bangs. Software developments

Two important developments in machine-vision software were also covered during the conference. One of these, a method for compensating for uneven lighting and poorly jigged parts, was described by Dr Stanley Lapidus, president of ITran Corp, Manchester, NH. The Itran development will be described in detail in an upcoming issue of Tooling & Production.

The other development, dubbed "mathematical morphology," was described by Dr Stanley Sternberg, president and chief technical director of Machine Vision International, Ann Arbor, MI. Now used by three suppliers of vision systems, mathematical morphology is said to provide higher speed and greater accuracy than other methods used for optical gaging.

Put simply, the method involves repeated dilation and contraction of images in order to extract important features. Dr Sternberg showed a series of slides that illustrated use of the method for inspecting for broken teeth on gears. Next year, bigger and better

According to John McEachran, conference manager with SME, next year's vision conference will be held in quarters that can accommodate a much larger show by vendors. "Along with the conference, we'll have a major exposition," he said. "Many more vendors will be able to display their systems, and visitors won't find the aisles so cramped."

Tentatively, the fourth annual Applied Machine Vision conference is scheduled for March 26-28, 1985, and is to be held in Cobo Hall, Detroit, MI.

RI/SME is offering bound copies of the proceedings from the third annual conference at $45 a copy to members of SME and its associations, $50 a copy to nonmembers. If you'd like to receive descriptive literature and an order form, circle E5.
COPYRIGHT 1984 Nelson Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1984 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Quinlan, Joseph C.
Publication:Tooling & Production
Date:Jun 1, 1984
Previous Article:Micros in manufacturing; personal computers in engineering and manufacturing.
Next Article:Aircraft parts maker automates fabrication.

Related Articles
Flexible assembly pays off.
Machine-vision systems - what can they do for you?
Assembly cell speeds riveting: workholding assembly tool serves family of aircraft structures for big payoffs in small lots.
Stanley Machining is new partner with ABB.
AVT installations.
Hyundai taps UNOVA IAS for lambda C-6 engine line.
RTS achieves the ultimate challenge.
Sting in the tail.
Flexi Pro Robotic Vision Picking Cell 'Adept' at fast robots.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters