CAMERA TRENDS 2013: A Simpler Machine Vision
- Created: Sunday, 01 September 2013
Today’s cameras are cheaper, smaller, and more capable than ever before. Without breaking the budget, a company manager who needs products inspected can buy a camera, quickly code in commands to recognize an object, bracket the device above a production line, and begin capturing images.
Machine vision customers are also expecting ease of use in production-line vision products, as well as the installation process. Users want high performance without a high requirement for expertise.
To put it simply: Simplicity.
As part of Imaging Technology’s OEM Camera Directory & Guide, we look at three technology factors that are simplifying the imaging process for customers: System integrators, smart cameras, and software.
The whole package of camera and machine vision technologies is getting more compact and more powerful. What used to be a cumbersome camera that needed a lot of wires, software, and sensors now only requires a small footprint of a device. As customers seek simplicity and ease of use, analysts have seen a continuing growth in a consolidated, more intelligent product: the smart camera.
A traditional camera requires a processor — usually a computer. With a smart camera, however, all of the processing is done in the unit itself. A typical smart camera has a lens on the front, along with an Ethernet port on the back, which can go straight into a user’s network.
The appeal of the smart camera is that it is cheaper and easier to use; the technology is hosted in one unit. No cables are needed because everything is connected internally. Thanks to low-power multicore processors, the simplified cameras also have improved megapixel capability, higher speeds, as well as plugand- play functionality. Customers can simply run an Ethernet cord, FireWire, or USB from their computers housing the program software.
“There are units that can easily fit into your hand that are actually very clever, which, at one time, took one or two people to pick up,” said John Morse, senior market analyst at the Wellingborough, England-based IMS Research.
If a customer needs to inspect a given product, he or she can prop a smart camera on a bracket, point it at the object, and, in effect, self-teach the device. Pressing a “Teach” button on many cameras, for example, will allow the technology to store a product in its memory, and enable the device to recognize products coming down the line that do not meet the programmed specification.
Alexander Shikany, director of market analysis at the Ann Arbor, MI-based Association for Advancing Automation (A3) — the umbrella association over Robotic Industries Association (RIA), AIA (Advancing Vision & Imaging), and Motion Control Association (MCA), which represents over 650 automation organizations — said that the smart cameras today are more powerful and simpler to use.
The smart camera architecture has a more compact volume compared to a PC-based vision system, and a simple — in some cases, absent — user interface. The consolidated architecture has led customers to use them for less demanding applications, such as inspection, said Shikany.
“In the machine vision industry, a smart camera may be likened to an iPad in the PC world,” he said. “While they are often used for simpler applications, advancements in technology have kept their hardware competitive with other machine vision systems in today’s market.”
Developments in imaging software have enabled the “self-teaching” capabilities of the smart camera. Machinery from manufacturers like Cognex, Microscan, or National Instruments, for example, allows users to point a camera at an object, press a button, and set parameters so that the machine can accept or reject it based on its characteristics. “Almost, kids can program it,” said Morse.
Slick software writers have come up with the software that allows users to generate 3D images, and check them against a standard shape, in order to determine a go or no-go product in an assembly line.
Improved software also enables new functions in robotics. Camera-equipped robots, for example, can twist and orientate themselves, recognize and pick up a lost object, and put it back neatly on an assembly line.
“The actual mechanics of these things are not particularly new, but the software to make it all happen reliably definitely is,” said Morse.
Shikany said imaging software has improved by leaps and bounds, mentioning a demo of National Instruments LabVIEW software package at the Vision 2012 show in Stuttgart, Germany. National Instruments vision development software includes IMAQ Vision, a library of vision functions, and IMAQ Vision Builder, an interactive environment for prototyping vision applications.
Instead of needing a substantial knowledge base like C++ to go into the software, change commands, and tweak functionalities, Shikany watched a user at the vision conference click a number of easily coded, easily readable buttons sequentially, and within a matter of 60-90 seconds, modify the programming for a USB3 camera. The attendee quickly changed the speed settings of the camera.
With the technology demonstrated at the conference, a user does not need to spend lots of time fidgeting with a camera’s settings. If a factory needs to increase production and run its line faster due to an increase in orders, for example, a user can make efficiency-driven changes within minutes.
“Software packages like that are allowing for this new wave of people to come and work in the industry because they’re able to easily learn how to run and implement the software,” he said.
Although cameras and machine vision tools are getting easier to use, customers still expect assistance in the implementation process. One answer for users with service needs is the system integrator — a company positioned between the manufacturer and the end user who is implementing a machine vision system. The system integrators provide support to the end users.
“Not only have the technology limitations raised expectations, with increases in technology, but as time has moved forward, the end user is expecting more support, more service, and more integration capabilities from these companies,” said Shikany, who sees a growing role for the system integrator.
The system integrator purchases the products, and often has partnerships with given manufacturers on a company-by- company basis. The integrator purchases the components from the manufacturers and then works with customers in project-based situations.
“So as you might expect, these system integration companies, over time, gather a lot of useful experience and knowledge,” he said. A3, Shikany’s organization, offers an AIA Certified System Integrator Program for system integration companies.
ATS Automation, a large system integrator based in Cambridge, Ontario, Canada, constructs production lines for various industries. The company, with offices in North America, Europe, and Asia, is divided into segments, including Life Sciences, Automotive and Transportation, and Energy and Consumer. The company works on production lines for specialized medical devices ranging from diabetes testers and syringes, to automotive devices and electronics.
A typical customer provides ATS with its specific vision system needs, including what the inspection is, what it needs to detect, and what the user requirement specifications are. The ATS team also receives sample parts from the customer; the parts help the crew’s engineers determine failure modes and specific measurements. Then, the team begins automating the machine-vision process.
Jay Stavnitzky, senior vision engineer at ATS, said customers are generally scaled to produce their product, but they lack the ability to automate on their own.
“In terms of coming up with new inspections or new automation even, quite often that’s out of their depth as a customer, and that’s not their focus. Their focus is to make money making whatever it is that they make,” he said.
Stavnitzky said the role of the imaging market systems integrator has changed. In previous years, there was much hesitation with imaging systems on production lines. If a camera broke, for example, a facility may not have someone with the expertise to fix the problem. Customers, he said, tended to solve their problems mechanically, because that is more in line with what the customer is able to support in-house.
With vision systems on production lines, however, the implementer is dealing with the mechanical and electrical, as well as computer and optics, which is where a system integrator, with a wider range of expertise, comes into play.
“That combination requires a sort of specific skill set to implement and maintain well,” said Stavnitzky. “Now, there’s more acceptance of vision in imaging on automation equipment.”
Easy-to-use technologies like USB3 and smart cameras, along with players like systems integrators, have all pointed to a customer’s demand for simplicity: maximum function with minimum expertise. Manufacturers are therefore pushing their products toward keeping it simple.
The imaging process is, in fact, getting easier. Small machine vision cameras from companies like AVT, Basler, and Point Grey sell for around $400. Developments in software have increased capabilities, often at the click of a button. Today’s imaging sensors are faster and cheaper, leading to faster and cheaper cameras.
“There are so many products out there that are very clever and very easy to use that unless you can actually meet that requirement, you’re going to fall behind,” said Morse.