
Machine vision provides defect-free aluminium containers
The Swiss company, Leuthold Mechanik AG (HLM) makes machines for manufacturing aluminium containers which, among other things, are used for packaging pet food and foodstuffs. A core element of quality assurance is machine vision by STEMMER IMAGING.
Video: Defect-free aluminium containers with machine vision from STEMMER IMAGING.
Unremittingly, 120 times a minute over three shifts the ram on the press in the HLM Competence Centre rises and falls and after every stroke spits out four finished pet food containers. A machine like this thus produces 480 high quality aluminium containers per minute, which will later be filled with pet food.
There are several reasons for using aluminium as the material for such containers: For one, this material is gas-tight, which means that the contents can be stored for much longer. Furthermore, aluminium can be recycled and consequently is more sustainable than plastic, which is not gas-tight and also cannot be classed as harmless from a health point of view, due to the addition of plasticisers. The aluminium cans are also coated, so that the foodstuffs do not have any direct contact with the aluminium.
“However our machines not only produce containers for pet food but also the most varied forms of aluminium receptacles for, among other things, preserves, pies or ground coffee”, explains Mathias Leuthold, who is responsible for the management of toolmaking and mechanical engineering in the family business. The company has also developed machines for entirely different substances such as fuel pastes or packaging for medical products such as inhalers.
Such machines are used only to a very small extent in the company’s own production facility in Einsiedeln and these are mainly for test purposes and for further development of the machines. A considerably larger proportion is installed with HLM’s customers worldwide. “In addition to the domestic Swiss market, our customers are based in many European countries. But our machines are also in use, e.g., in the USA and Australia”, says Mathias Leuthold.
The prerequisite for stamping out thin-walled containers, the dimensions of which can vary between 60 and 120 mm in width and between 60 and 200 mm in length, is high quality aluminium foil. The stamped-out containers are conveyed in line to the module-based built-in inspection and stacking machines.
“If a customer has decided on us as his machine supplier, the first step as a rule is to send us some ideas, sketches and/or 3D models of his desired aluminium container. On this basis we will then develop the whole machine, from the design of the container or casing through the tool and checking module to the stacking unit, or adapt our standard machines to the customer’s wishes”, explains the engineer describing the usual path from the idea to the finished machine.
Machine vision as the core element
After the containers have been brought into the finished shape by a single stroke in the press, they are blown out and transported in-line by means of transport belts and mechanical tracks to the modules where quality control is carried out.
“Aluminium is relatively expensive as a base material, so the walls of the containers must be as thin as possible to keep the costs low. On the other hand, as the material thickness is reduced, there is an increased risk that holes will appear during the forming process as a result of inclusions in the raw material or excessive stresses during the forming. These holes will cause the container to leak and thus make it unusable”, says Mathias Leuthold, explaining why every single container must be checked. “We must detect and separate out every defective container as otherwise, using the example of pet food or foodstuffs in general, there is a risk that the contents will go bad.”
Due to the high production speeds and the required 100% inspection, machine vision is a viable tool for quality inspection. HLM has already been using this technology for over 20 years and has had positive experiences with it, Mathias Leuthold confirmed: “There are no longer any installations in which machine vision is not used as a central element of the testing stations.”
Changeover to Embedded PC systems
For the latest generation of machines, the Swiss company is no longer relying on intelligent camera systems as before, but rather on an Embedded PC vision system. According to Leuthold, several benefits justified this decision. “With the previous architecture using intelligent cameras, a PC first had to be connected to the machine in order to display images of defects or undertake statistical evaluations. Only in this way was it possible to check, for example, whether more defects arose on one of the lines or whether a certain type of defect occurred more frequently. The operator can now detect such tendencies much more easily and more quickly via a monitor directly on the machine, and thus switch off the defect sources selectively in a shorter time.”
This facility is implemented by means of a ring buffer, which saves the last 20 defect images in the embedded system and displays them on demand. In addition to images of current results, the user can also view statistics about types of defects and their frequency, the distribution of rejects across the different tracks and also images of defective containers very easily, directly on the machine’s display. “This contributes significantly to faster rectification of defects by means of appropriate mechanical adjustments to the machine”, emphasises Mathias Leuthold.
As a further advantage of the new embedded system compared with the previous intelligent cameras, he cites the improved connection to the users MES (Manufacturing Execution System) systems: “This enables optimised control of production and clear advantages for data acquisition.”
From an economics point of view, the costs of the earlier intelligent camera systems were roughly on the same level as those of the embedded PC systems of the new generation of machines, however the additional benefits tipped the scales in favour of the embedded PC based solution.

Expert machine vision knowledge
HLM does not employ its own machine vision specialists. “Although we have already been using this technology for around 20 years and have accumulated some experience in this matter, we have nevertheless been relying on the know-how and informed advice of our partner, STEMMER IMAGING, for many years", Mathias Leuthold pointed out.
According to his statement this also applies particularly to the recently implemented changeover to embedded PC-based systems, where the Swiss subsidiary of STEMMER IMAGING was concerned primarily with the optimum selection of all machine vision components. A feasibility study then led to a recommendation of which imaging products are best suited to the new generation of machines.
“We have checked and then approved this selection, and are highly satisfied with the results from the first machine”, enthuses Mathias Leuthold about the successful collaboration. Furthermore, he highlights one service in particular: “STEMMER IMAGING has delivered all components to us already pre-configured and tested, which has considerably reduced our expenditure in implementing the machine.”
A LED light is mounted on the cover of a testing station, shown here opened.
In the new generation of machines, there is a Genie Nano camera, made by Teledyne DALSA, installed on every line. This is triggered by light barriers. Equipped with suitable optics from Lensation, these cameras are located underneath the transport tracks. The deciding factor for selecting the Genie Nano was the fact that these cameras met all the requirements with respect to performance, resolution and speed and furthermore were attractive from a price point of view.
The LED lighting, which is also triggered by light barriers, is integrated above the testing stations in the machines, so that the containers can be inspected in the transmitted light procedure. It was developed by STEMMER IMAGING specially for the present requirements. A certain size and a defined lighting angle are required for this application. “There is no standard lighting that meets these requirements, so in this case we have developed customer-specific lighting for Leuthold Mechanik”, says Claudio Sager, Managing Director of the Swiss subsidiary of STEMMER IMAGING, to justify this step.
The embedded PC, which undertakes the evaluation of all the images from the four lines and also handles the control of the machine by means of the so-called Real Time Manager, is a custom product from the Swiss electronics manufacturer, Worx. It has to manage a considerable amount of computing power. With this machine it must evaluate 120 images per minute and track, and forward the results to the discharge station in the shortest time possible so that defective containers can be separated out immediately. The images are evaluated using the Common Vision Blox (CVB) imaging software from STEMMER IMAGING.
“Even with this central component, the applications know-how of STEMMER IMAGING was decisive in finding the perfect solution for us”, Mathias Leuthold praises the collaboration with his machine vision partner. According to him, good cooperation with Worx, where the installation of the software and the programming of the embedded PC were carried out exactly in accordance with Leuthold Mechanik’s specifications, was also important.
Currently the engineer and his colleagues are completing the second machine of the new PC-based generation which, after training in machine operation, will start production in a few days at a customer’s premises in Austria. “And many more will follow”, Leuthold is certain.