Machine vision in the packaging industry
What are the specific challenges for quality assurance in the packaging industry?
Gillig: The diversity of materials and shapes used in the packaging industry regularly provides us with new tasks. Packaging has many functions: it protects raw material during transport but also positions a product on the shelves in supermarkets and the materials used are changing all the time. For example, foils are often manufactured in multiple layers and then coated which makes inspection of the goods more difficult. Food or cosmetic products differ considerably in their packaging. Furthermore, the shapes and colours of a product are often adapted seasonally. This multiplicity of factors places special demands on imaging, as every inspection station which checks a product via image processing for faults needs to be adapted to the new conditions for every new material and every newly designed product.
What are the actual fields of application for machine vision in the packaging industry?
Gillig: Here we have a host of potential applications. For example, the quality of the raw materials for packaging is inspected using machine vision: is a foil supplied free of creases and wrinkles or the smallest of holes? Do the bottle walls of injection-moulded plastic bottles have streaks? Has the label of a container been designed counterfeit-proof and how can this be verified?
Additionally we can also check the dimensions or shape of an intermediate product during the production process with vision systems. A good example is checking whether punching or cutting of cartons is done properly or if the thread or shoulder of a glass bottle have been formed correctly. Reliable reading of embossed numbers of bottle bases, the inspection of cans and sealing caps for true-sided assembly, or the consistent checking of expiry dates on colour-printed aluminium caps are just some of the examples for applications in this field.
At the end of the production process labels often need to be checked. Barcode labels are affixed at different heights depending on the product or are already foil-sealed. These cases also require careful selection of the imaging components and optimal matching depending on the task.
Which STEMMER IMAGING products are used in the packaging industry and does STEMMER IMAGING offer special solutions for this industry sector?
Gillig: As the largest provider of imaging technology in Europe, STEMMER IMAGING offers its customers all the components necessary to solve machine vision tasks. Our product portfolio holds everything necessary for a suitable vision solution and includes a large selection of lighting equipment and optics for a broad range of industrial cameras of all types, right through to intelligent camera systems, high performance image capturing products, complete systems, the requisite accessories and suitable software.
The choice of the right lighting is often the decisive aspect for the subsequent performance of the inspection station, as every search for faults or inspection of shapes requires an optimally matched lighting design. It must be able to highlight the characteristics of a material with adequate contrast and suppress or pronounce targeted gloss or reflections. In some applications it also makes sense to highlight or level colours. Lighting makes the image, which is why we often develop the optimal set-up for the respective task in feasibility studies and then offer the customer this solution.
Of course the camera used also plays a major role. Area or line cameras with one or more sensors may prove to be the optimal solution here, for example, to distinguish colours. Depending on the task, special prism cameras may also prove to be the right solution for evaluating the red, blue, green or IR ratio in the light spectrum in order to display colours more clearly and reliably.
At this point I would also like to mention the new contact image sensors (CIS) by Mitsubishi Electric as an example of a complete system for the packaging and printing industry. These products operate like scanners which capture flat materials at a resolution of 600 dpi and a fixed operating distance. As the lighting is already integrated in the CIS housing and the sensor is easy to mount, these systems are very suited for inspecting flat objects such as foils, paper, textile sheeting or glass. However, CIS technology is not suited for the inspection of uneven surfaces such as punched cartons.
What should a company in the packaging industry look out for when choosing a machine vision system?
Gillig: I would not recommend completely pre-configured systems. The components employed, in other words, the camera, lens, lighting and software and possible other elements need to interact optimally to guarantee performance. For example, pre-configured systems often have pre-installed lighting which usually does not meet the requirements of the task in question and therefore suitable lighting needs to be retrofitted at great expense. This is usually difficult mechanically and the price advantage compared with an individually configured system is soon lost.
Another important point is to allow sufficient time from the beginning for selecting the right components, especially the lighting. With increasing experience one gets a feeling for which components represent a suitable solution depending on the material to be inspected and other framework conditions. Often our customer’s employees enter new territories with machine vision. To reduce familiarisation times in these cases and to obtain the desired results faster, we offer regular training courses on all topics of imaging and machine vision, and these are rated very positively by participants.
How does a machine vision system pay for itself in the packaging industry (i.e. process acceleration)?
Gillig: Machine vision does not make machines run faster and especially in the packaging industry, the packaging machines and production processes are already extremely fast. However, the machine vision industry offers its users cameras which can capture images for evaluation at these high speeds, as well as frame grabbers which allow pre-processing of images, so that image evaluation takes place in a few ms. Suitable frame grabbers transfer the the result direct to the I/O's and thus provide reliable rejection of faulty units. Thus suitable systems reduce the load on the CPU of the connected computers.
Machine vision can deliver sizeable savings if inspection which was previously done manually or partly automated can be further automated. It is my opinion that there is still tremendous potential for this technology in the packaging industry.