Early Obstacles
Cognex received one of its first patents for Search, a powerful
software tool that dramatically improved machine vision
performance by enabling quick, accurate location of patterns
in gray-scale images. Cognex now has a portfolio of more than
250 patents for advances in machine vision technology.
Despite a growing list of customers using Cognex vision, serious challenges soon became apparent in the company’s business model. Implementing a vision application in these early years required computer programming knowledge, and users demanded considerable support.
In addition, factory conditions were unpredictable. Early systems often did not perform reliably outside of development laboratories, where factors like lighting, reflections and shadows could be controlled. Misperceptions about what vision could realistically achieve combined with poor reliability in factory settings resulted in many vision companies spending too much time supporting customer applications. As a result, Cognex…and every other company that had entered the vision business… was losing money.
In 1986, Cognex made a major technical breakthrough that helped solve the problem of system reliability. Cognex co-founder Bill Silver developed a powerful software tool called Search that could locate patterns in gray scale images very quickly and accurately, and succeeded in dramatically improving the results that users could achieve with their vision systems.
At the same time, Cognex launched a new business strategy that helped ensure the company’s survival and subsequent market leadership.
That strategy was to develop and sell standard machine vision hardware and software products to original equipment manufacturers (OEMs), who could integrate machine vision into the manufacturing equipment they sold into factories. These OEMs had engineers on staff with the expertise to program the vision applications, and who then supplied end-users with equipment that had the vision already built in.
This combination of superior technology and new business direction provided the winning recipe that would help Cognex succeed and grow, while the majority of other early vision companies soon failed or got out of the business.
First Vision System
Cognex’s first industrial OCR reader, DataMan. Cognex
founders celebrated with a champagne toast when the first
DataMan prototype system successfully read its first character.
It took the system 90 seconds to read the number “6”.
The company produced its first vision system, DataMan, in 1982. DataMan was the world’s first industrial optical character recognition (OCR) system capable of reading, verifying, and assuring the quality of letters, numbers, and symbols marked directly on parts and components. Cognex’s first customer was a typewriter manufacturer, who purchased the system to inspect the keys on each typewriter to ensure that they were located in the correct position.
Cognex was one of the earliest companies in a market that was soon crowded with competitors, all intent on securing a position in the new field of machine vision. In these early years, machine vision generated great excitement as part of the “robot revolution.” People believed that machine vision would revolutionize not only manufacturing, but even areas as diverse as transportation and household chores. The reality was not to be as easy…or come as soon….as predicted.
What are Vision Systems?
Vision sensors are cameras with and imaging electronics in the form of CCD or CMOS chips. They are used in quality control, inspection and design and manufacturing testing applications. Vision Systems often include interfaces for PLC programming and data output, and also have a variety of measurement and inspection functions.
When specifying vision systems, it is important to determine whether a monochrome or color sensor is needed. Monochrome vision systems present the image in black and white, or grayscale. Color sensing vision systems read the spectrum range using varying combinations of different discrete colors. Multiple chip color is found on some vision systems. It is a method of capturing color in which multiple chips are each dedicated to capturing part of the color image, such as one color, and the results are combined to generate the full color image.
Some important considerations when choosing vision systems include number of images the vision system can store and maximum inspection rate. The number of images stored represents captured images that can be stored in the vision camera. The maximum inspection rate is the maximum number of parts that can be inspected or evaluated. This is usually given in units of inspections per second. Other important parameters include sensitivity, frame rate, signal to noise ratio, horizontal resolution and shutter speed.
Inspection include pattern detection, edge finding, image direction, alignment, object measurement, object position, bar or matrix code, optical character recognition (OCR), and color mark or color recognition.
Other parameters to consider when specifying vision sensors include performance features, physical features, lens mounting, shutter control, sensor specifications, dimensions, and operating environment parameters.
Vision Systems Diagram
Vision Systems Diagram
