Bin Picking with EVT’s Deep Learning Recognizer
Safe Bin Picking is the combination of precise artificial intelligence recognition of a correct position in space with 3D and robot hand eye calibration. These functions are combined in the EyeVision software and enable fast and easy bin picking.
Even demanding applications such as loose, stacked cartons or bags lying on top of each other etc. are reliably detected and the correct gripping data is transmitted to the robot via the 3D data. The entire sequence can be put together quickly and easily using drag-and-drop, after which the robot works autonomously.
The Hand Eye calibration enables the correction of the absolute accuracy. As a result, the robot can grip even the smallest components safely and repeatedly in the entire installation space. The correction values are determined automatically and with minimal effort.
The multi-camera support of the EyeVision not only enables precise gripping, but also the placement of the components at the target with camera support. This means that even with different components (e.g. cartons), they can be stacked in the appropriate spaces.
By supporting all camera types of the EyeVision software (1D, 2D, 3D and thermal cameras), the appropriate camera can be selected for the task to be solved. For example, a 3D camera to remove the components from the box and a 2D camera to place them in another location.
As always, the software is integrated completely graphically and for communication with the robot and all other systems, a variety of interfaces are available, including OPC UA, Profinet, TCP IP and others.