Vision sensing technology in track-and-trace applications

Sensors work with control systems to follow perishable and critical products through manufacturing and distribution.

By Mark Lampert, Steve Wong December 7, 2011

Vision sensors are a critical component to an industrial manufacturing process, helping to optimize productivity and improve product quality by detecting faulty parts before they leave the production line. Vision sensors see, analyze, and interpret data in complex visual inspections, including parts verification, gauging, measuring, orientation, flaw detection, and sorting. Vision sensing is especially significant in food manufacturing processes, providing track, trace, and inspection functions to follow a product from initial development to the final packaged result.

Vision sensing (or electronic imaging) provides control to manufacturers over multiple processes from machine tools and robots to product quality and inspection. Starting with a camera, which determines the range of the product features and performance, and a control system that is used to interface the sensors with machine controls, vision sensors allow users to set product-specific parameters for inspection. The inspection process consists of a three-step process: The camera acquires an image of the part; next a processor analyzes the image; and finally, it determines if the inspection passes or fails, reporting the results to the manufacturing line.

Vision sensors are a superior manufacturing detection tool due to their unparalleled abilities to handle high-speed assembly lines, allowing the sensor to inspect products with 100% accuracy. This is in comparison to human inspection capabilities, which at best are approximately 80% accurate. Additionally, vision sensors are able to conduct simultaneous, repetitive, multiple, and consistent inspections on objects (same or differing), providing reliable results for complex detection needs. With their user-friendly, easy-to-install application, enhanced functionality, plus reliable communication to operation devices, today’s vision sensors offer a nearly universal vision solution, delivering cost-effective, powerful, and controlled vision capabilities for manufacturers.

Beyond presence detection

Vision sensors are inspection tools using and interpreting light (or the absence of it) to control machines and processes for product detection. Unlike traditional photoelectric sensors’ simplistic operation of presence detection, vision sensors can also analyze the color, shape, size, or position of an object’s feature. Vision sensors do not rely on the object to trigger inspection, but instead use dedicated light to illuminate and capture the entire image for complete visual feedback. A single vision sensor can inspect multiple points on an object, reducing equipment requirements while gaining expanded range. The vision sensor “sees” an area, and senses many thousands of distinct points (or pixels) in the specified area. Further, users can change the lens of the sensor, making the field of view smaller or larger, to obtain accurate detail, even at a microscopic level if needed. Upon image capture, the vision sensor provides a binary output with enhanced communication capabilities, allowing users to send large amounts of data about an object’s parts or features serially or via computer protocols such as Ethernet. This enables data to be communicated to the entire manufacturing enterprise.

For example, in food manufacturing processes, a vision sensor can verify the correct label on the food product’s packaging by reading the 2D Datamatrix barcode or matching the pattern on the label. This not only prevents incorrect cartons or packaging from being added to a product, but also effectively tracks and traces the product throughout the entire manufacturing process. In addition, a vision sensor can verify that the date and lot code are printed legibly, ensuring the correct range of products are easily identified if a recall occurs. 

Enhancing communication

Most vision sensors have multiple communication methods to interface with the control system, ranging from simple discrete I/O to serial communication to Ethernet. Discrete I/O is used for parts of the system that require higher speeds or simple controls, such as trigger or pass-and-fail outputs. Serial and Ethernet communications are used for data transfer and system information, including pattern names, pass/fail counts, and program selection. When high-speed communication is necessary, such as when exporting images, Ethernet is the primary method.

Ethernet connectivity provides ease of configuration and networking capabilities in manufacturing processes. With the advent of standard industrial protocols over Ethernet, vision sensors have the ability to exchange input and output data directly to PLCS, HMIs, or alternative factory devices with minimal programming of commands, allowing operators to retrieve data quickly and easily. For example, users can access valuable inspection data such as inspection time, pass/fail status, system error, trigger (or mis-trigger), ready signal, and additional inspection data directly from the factory device. Additionally, data used for authentication can be remotely accessed and updated, including up-to-the-minute pattern variations providing unparalleled control over production. In barcode applications, Ethernet communication also permits the reference barcode value to be uploaded automatically into the sensor.

Improving inspection

Recent advancements have improved the ease of use of vision sensors. Advances in touchscreen capabilities and inspection algorithms have allowed sophisticated sensors to break free from the PC, eliminating the need for cables. For example, some vision sensors feature inspection tools like match, blemish, sort, and barcode readings that are powerful, yet simple to configure. Installation and configuration is easier using menu-driven tools that guide users through selecting specific inspection designations for each product. This allows a plant technician to configure an inspection step to sort between multiple products without the need to bring a laptop to the manufacturing line. With some advanced vision sensors capable of storing up to 100 inspections at a time, working with high-speed assembly lines and rapid product changeover is simple. Configuration and testing can be done remotely using a sensor emulator, which allows offline updates and modifications, shortening setup time on actual production equipment.

Vision sensors now offer remote monitoring possibilities, pairing a sensor and separate touch screen to support inspection capabilities in difficult-to-monitor areas. Once vision sensors have been installed, users can operate the remote features to configure, monitor, and update inspection requirements for all vision sensors, providing a new comprehensiveness to product inspection. 

For food manufacturing and other washdown environments, vision sensors can be provided with enclosures of IP67 ratings or greater in plastic or stainless steel to accommodate harsh conditions.

Field applications

Vision sensors can perform specialized functions in food packaging and distribution applications, ranging from reading barcodes on cartons to verifying packaging labels. For example, many food operations require a date/lot code to be present on the packaging. To verify each package has that information printed on it, a vision sensor can use a match inspection function, which captures an image of the date/lot code. When the inspection is running, if the sensor detects a package without the code, the sensor sends a fail output to the control system, which tells the line to reject the project. This ensures the correct range of products is easily identified if a recall occurs. 

Track-and-trace functions are also crucial for ensuring product authenticity. For instance, a vision sensor can effectively inspect and authenticate wine bottles through barcode readings. A vision sensor first reads the 2D Datamatrix codes, typically printed on the bottles’ labels. The scanned information is then either inspected by the vision sensor’s internal data comparison tool or sent externally to a remote secure database. If a barcode is misapplied for that product, is not recognized by the system, or suggests a possible counterfeit, the product is rejected and quarantined for investigation. Vision sensors configured with 2D Datamatrix barcode reading capability are commonly used to read and verify 2D Datamatrix barcodes to track and trace products in the manufacturing process and throughout the distribution infrastructure.  

As vision sensing technology continues to develop, manufacturing processes will experience significantly improved error-proofing and productivity, creating a product that is not only of the highest quality, but also meets the most critical industry expectations.

Mark Lampert and Steve Wong are business development managers for Banner Engineering.