Sorting food on packaging lines with 3D localisation

The ability to ensure hygienic primary packaging of sensitive open food is vital within the food industry. This is why SICK is working with robot manufacturer Stäubli to make this process possible within a robot cell. The TriSpectorP1000 3D vision sensor from SICK and the four axis FAST picker TP80 HE from Stäubli solve picking processes while the conveyor belt is in motion. 3D parts localisation is implemented by the Belt Pick SensorApp and SICK AppSpace. In case studies prawns or biscuits were successfully sorted in packaging trays.

Both the food and pharmaceutical industries have extensive requirements when it comes to sorting and orientating small objects in a hygienic packaging process. By working in partnership with Stäubli on a pilot project to sort prawns using the FAST picker TP80 HE while the conveyor belt is in motion together with the programmable TriSpectorP1000, SICK is opening up a diverse range of applications. From salad to biscuits to sausages, the TriSpectorP1000 is capable of handling a whole host of different tasks.

The 3D technology is suitable for objects of irregular size and variable height, as well as for double layer detection. Together with an encoder, the sensor is capable of detecting each individual object accurately and reliably. Once the position of an item has been recorded on a three-dimensional basis, the exact coordinates are passed on to the FAST picker TP80 HE. Its vacuum gripper then removes the prawn from the belt one by one and places them in the trays according to a specified packing pattern. The four axis robot arm can typically handle over 200 picks per minute. This particular application does not require maximum speed, yet the robot is still capable of sorting over 60 prawns a minute.

The TriSpectorP1000 records the position of the products by means of a laser triangulation process, which also calculates their height and volume. This is particularly significant when packaging organic products, as their sizes tend to vary. Furthermore, the ability to calculate volume makes it possible to reject prawns that are too small in this particular application. The FAST picker TP80 HE also uses this measurement data to achieve a visually appealing presentation within the tray and ensure a standardised packaging weight.

The future looks bright: tailor-made 3D vision solutions with SICK AppSpace

The robot’s 3D localisation capability enables it to detect the spatial position of an object as part of a sorting task, regardless of the contrast situation. It can even prevent collisions with the object thereby avoiding any resulting damage. SICK offers tailor-made 3D vision solutions for flexible automation: SICK AppSpace is an eco-system of software tools and an expanding number of programmable devices based on technologies such as 2D and 3D vision, RFID and LiDAR. By using the application programming interface (API) functions, so-called SensorApps are developed.

The SICK Belt Pick Toolkit is a SensorApp specialised for locating products on a conveyor belt which is also in use here in cooperation with Stäubli. With this piece of software installed, the camera turns into a 3D belt-picking sensor. This also offers excellent prospects for the future: No matter whether the task involves sausages or medication, product changeovers require virtually no adjustments as long as the mechanical dimensions are similar. Conversions such as these couldn’t be easier with the TriSpectorP1000, as they are all handled automatically and digitally using SICK AppSpace to make the entire process of handling packaging lines more efficient all round.

Image credit: SICK