SICK and ABB robot picking solution with 3D vision

Click2Contact

ABB and SICK have developed a robot picking solution using 3D vision technology, in order to deal with objects of variable size and thickness arriving on belts in random positions.

Conventional 2D vision systems lack the height and volume information for determining accurate position when goods naturally vary in terms of height and shape. 2D vision systems determine objects using contrast and colour, while a 3D detection system is based on height and volume measurement. This means that objects can be detected more precisely in situations where the contrast between the objects and background is low or varying, the objects vary in height and shape, or when the carrier belt is wide enough to cause position distortions in a 2D system.

Detecting the height of the object increases throughput and reduces the risk of damage. Using this technique, a robot is able to pick up objects more quickly, with greater options for volume and weight inspection and detecting anomalies on the object.

The robot utilises SICK‘s IVC-3D vision sensor and the ‘belt picking toolkit’, a piece of software which allows the system to carry out height-based object detection. Based on laser triangulation, the IVC-3D can capture and report data in three dimensions.

Data being reported contains both three-dimensional position data and timing information — the combination of these two data sets allows ABB to control the picking process in both 3D space and time. ABB’s fast IRB 360 FlexPicker robot and high-performance line master controls are most commonly used.

The system is suitable for products requiring gentle handling, such as pasta packed in pillow bags, fresh cheese, nuts, frozen berries and vegetables, bread, fresh fruit and vegetables, and meat.