Create a free account to continue

Visual Aids: Efficiency You Can See

Machine vision technology can help improve efficiency in food manufacture and packaging.   Many manufacturers already benefit from the use of machine vision in food production and packaging. The technology is found in a variety of applications in this industry ranging from quality control to robotic guidance.

Machine vision technology can help improve efficiency in food manufacture and packaging.

Many manufacturers already benefit from the use of machine vision in food production and packaging. The technology is found in a variety of applications in this industry ranging from quality control to robotic guidance. With advances in the capability of general-purpose components and reductions in cost, machine vision is more accessible than ever.

Technology & components

One way of categorizing the machine vision systems in the food industry is to group systems as either "application-specific" or as "general-purpose." An application-specific machine vision (ASMV) solution is a complete system that is fully integrated and offered to perform a specific inspection task. An example of this type of device is the familiar high-speed sorter, which can differentiate good product from bad in sorting fruits, produce, and even grains. Other ASMV solutions exist for other tasks, and overall, if a competent ASMV device exists for a given task, it makes sense to consider that inspection product.

General-purpose systems are comprised of off-the-shelf components, including smart cameras and PC-based systems, which are custom integrated by the end-user or an outside systems integrator to perform the targeted inspection. The huge expansion over the past several years in the availability, variety, and capability of these components has fueled a growth in the use of machine vision inspection in food processing and packaging—making general-purpose offerings more accessible and sophisticated to suit a broad variety of applications.

Performing the inspection

A machine vision system can acquire an image upon demand and can be configured (or programmed) to perform extensive analysis of that image, in order to extract useful data about the object being inspected. The image may be monochrome or full color, although a grayscale image often can be processed more quickly and (usually) at higher resolution with respect to cost. A machine vision system does not just "take a picture" of a good part then compare subsequent part images to that picture. Although this analysis technique is one capability of some systems, most machine vision image features are extracted through the recognition and processing of individual geometric objects in an image.

General-purpose components include tools that perform edge extraction, contrast measurement, blob analysis, pattern matching, and many others. Image analysis tools range from simple to complex, and they are usually combined to form an inspection process suitable for the target application. Machine vision components are intended to interface easily with other automation components via discrete I/O signals, Ethernet, DeviceNetâ„¢, and/or serial in order to communicate results and data for process control.

Machine vision is very successfully applied in common food processing and packaging tasks such as raw product quality analysis and sorting, guidance for robotic packaging or other automated processes, package integrity verification, and package and label quality and content validation. All of these applications may be achieved with general-purpose components.

Keys to success

Proper image acquisition is absolutely critical to any machine vision application. The camera's ability to capture a correctly illuminated picture of the object to be inspected, with that object correctly presented at a fixed orientation with respect to the camera, is considered to be 80 percent or greater factor in the success and reliability of the inspection.

Key considerations when evaluating imaging and lighting for applications in food process and packaging include:

It is important to take into account other implementation issues that may affect the success of the overall application. Almost all food process applications require specific environmental protection. This includes providing washdown enclosures for all components. Furthermore, the enclosures often must be resistant to caustic and/or corrosive chemicals and liquids.

Finally, in working with food, users must understand the random and variable nature of the product. Unlike man-made products, food products have widely varying features, geometries, and structure. These process variations present challenges to automated imaging that must be overcome with proper lighting, imaging and processing tools.

Integrating it all together

Integration is the part of the project where someone must make everything work. Quite often, the integration of the inspection system into the automated process proves to be the biggest challenge of an application. Pitfalls may be avoided by diligently undertaking an application analysis and preparing a detailed specification of system expectations before any components are selected or work is performed. The result will be a more smoothly integrated, successful machine vision project.

In one application, Aptúra Machine Vision Solutions (Lansing, MI) utilized intelligent camera technology for in-process inspection of pickle packaging. A leading food packager required verification of an automated pack-out process for pickles. Pickle slices are automatically placed into a bottle, and in a secondary process, additional product is added. If the slices were incorrectly placed in the first process, the secondary process would fail, causing backups and downtime.

The customer implemented machine vision technology to perform an automated inspection to verify that the pickle slices were correctly positioned for the secondary pack-out process. A reject mechanism was used push any rejected bottles off of the conveyor without damage. The automated inspection significantly improved process efficiency by reducing return and re-work of incorrectly filled bottles, and by capturing the incorrectly packed product for correction.

One application challenge was that, as the pickles were wet and shiny, differentiation between product and bottle was difficult. In addition, all components used needed to be NEMA 4X, stainless steel suitable for a caustic and washdown environment.

Aptúra Machine Vision Solutions provided the customer with a turnkey inspection station featuring a PPT VISION (Bloomington, MN) IMPACT™ C-series machine vision intelligent processor and standard resolution camera. Two LED bar lights were located on either side of the bottle opening, thereby illuminating only the tops of the product. By utilizing a red monochrome LED, the green and white surface of the product was easily highlighted. Adjustable camera and lighting mounts and an air-driven reject mechanism were mounted to an existing conveyor. An operator interface console with flat-panel display was pedestal-mounted near the inspection station. A Control Panel Manager (CPM) interface provided the customer with product counts, analysis of images and inspection set-up access.

As demonstrated above, machine vision technology can yield significant improvements in efficiency, and it is widely applied in a variety of applications in food manufacture and packaging. To implement a successful inspection, users must gain an understanding of the technology, fully analyze and specify an application, and implement the proper tools and components. Where applicable, manufacturers should partner with a competent integrator or manufacturer with proven capability in the type of inspection to be implemented to develop the best vision solution for their application needs.