Smart Cameras Add Security, Capability to Industrial Applications

Launched 15 years ago, smart cameras experienced a quick rise in market acceptance thanks to their small size, common space constraints of adding machine vision systems to production lines, and competitive price point against PC-host systems. Then PC prices plummeted as volumes grew, and smart camera growth — with its premium price point — hiccupped although continued to increase.

Today, new industry drivers are helping swing the pendulum back. Smart cameras have more capability, greater security, and ruggedness compared to PC host systems. They also help businesses comply with new traceability regulations, offer seamless scalability to PC platforms as line throughput grows, and help engineers manage quality across a geographically diverse global manufacturing network.

Adding Security to an Insecure World

“Pharmaceutical companies love smart cameras,” says Rick Roszkowski, Senior Director of Marketing, Vision Systems Business Unit, at Cognex (Natick, Massachusets). “There’s more intelligence and security going into smart cameras in terms of logging credentials and tracking changes and electronic signatures. They want to know who made what change to what system and why. With the broader adoption of Ethernet, and bridging the camera to the PLC, file servers, and other peripherals, people’s PCs have network connectivity to the factory floor, and that means more need for security across the board.

“On top of that, you have regulatory agencies driving compliance requirements, and that’s not just in the pharmaceutical industry anymore,” he continued. “[Track and Trace] compliance requirements extend to food and beverage, and even automotive and other large-scale manufacturing industries with global footprints. While the pharmaceutical company needs to know who made a change to their systems for security purposes, automotive production lines may use a part from three different suppliers in different locations around the world. Different finishes mean you have to tweak the vision system, and you want to know how much change was made. Smart cameras’ connectivity and security features help make this a reality.”

Connectivity is one of the growing differentiators for smart camera providers. Microscan Systems Inc. (Renton, Washington) recently introduced Microscan Link, which allows results and set points on their smart cameras to be connected to industrial controls and networks over ProfiNet I/O, Ethernet/IP, and TCP/IP. Cognex also has moved in this this direction, offering interfaces to a number of fieldbus and industrial network protocols. Datalogic (Bologna, Italy) also is moving in this direction with its A30 and T4x Series. “Today, it’s critical that the smart camera can communicate with all other types of automation out there from robots to PLCs,” explains Bradley Weber, Director of Application Engineering at Datalogic Machine Vision Business Unit. “This also allows our customers to access any smart camera running Impact software via web browser without a client on the local PC or tablet. You need the client to make programming changes, but in terms of visibility, you can see the entire network without special software as long as you have the proper credentials.”

Connectivity isn’t the only feature blurring the lines between smart cameras and other automation systems, including PC-host machine vision systems. Software also is helping blur the lines between different platforms, according to Nico Hooiveld, Product Manager for Microscan’s Machine Vision unit. For example, Microscan’s Visionscape software is optimized for Texas Instruments’ DSP engine on the Vision HAWK smart camera but also runs on PCs that use standard CPUs. Datalogic’s Impact software suite, for one, runs on both the smart camera lines as well as the MX series of embedded vision processors.

While that continuity can be helpful for companies wanting maximum scalability as manufacturing requirements change, there are benefits to optimizing a software program around the specific needs of a particular platform. “If you’re willing to invest to understand DSP instruction sets, you can really tune and drive more performance using a DSP,” Roszkowski says. It’s a lot of hard work really optimizing inner loops and writing machine code. With smart cameras, you can’t simply go the newest, most powerful CPU because thermal dissipation becomes a problem for smart cameras; you have to maintain the small footprint.” Cognex offers In-Sight for smart cameras and VisionPro for PC-host systems.

Similarity Doesn’t Mean Sameness

Despite some similarities among smart camera processing platforms and connectivity helping to erase traditional barriers between different automation platforms, including other vision systems, customers have more options than ever when it comes to smart cameras.

Most smart camera suppliers offer a wide portfolio from “vision sensors” for simple presence/absence tasks to 5 megapixel smart cameras running full image processing software suites. Microscan offers a 2:1 ratio sensor designed for code and OCR reading, while Cognex will soon introduce a VGA-resolution sensor capable of 200 fps. Datalogic has consolidated on CCDs away from CMOS to help keep costs low thanks to volume savings with the ID line of code readers. “Between our A30 and T4x, the difference is about how intensive your machine vision program is and how fast the line is moving,” says Weber. “In the future, we expect more of our smart cameras will come with application-specific software wrappers, making them easier to use and faster to set up.”

“In due course, the borders between barcode readers, vision sensors, smart cameras, and PC based machine vision systems will disappear, and instead we will see a continuum of devices to match price/performance requirements for any application ranging from simple presence/absence applications up to complex gauging jobs,” concludes Microscan’s Hooiveld. “The scalability that Microscan provides across the machine vision product lines and the field upgradeability of the functionality for the AutoVISION and Visionscape software suites through license key codes is a good example of where I see the market going.”

Whether the customer opts for a vision sensor or smart camera running at 200 fps for high-speed inspection, one thing every customer needs to think about is the total cost of ownership for any machine vision solution. “So many users want to focus on the capital cost of the equipment and fail to appreciate the continuum of cost across the entire lifecycle,” explains Cognex’s Roszkowski. “Smart cameras have proven — and continue to prove — that they are a key tool for managing the total cost of an automation solution. There’s a place and a need for PC host solutions in factory automation — like when you need a 20 MP camera, or eight cameras on a relatively slow line and security isn’t a major concern. PC hosts have more processing power and RAM than you can get on a smart camera. But to overlook the size, security, ruggedness, and compliance that comes with a smart camera solution can be much more costly than the acquisition price.”