The Visionary-T 3D snapshot sensor from Sick enables robots for goods separation to optically capture the parts. 3D snapshot refers to an image capture process in which cameras not only capture a complete image of their field of view in a single moment, but also determine the distance between the camera and the captured objects. The main advantages of this process include the complete and sharp representation of any object and the high depth resolution of the 3D images, which is particularly beneficial for precise robot guidance. In addition to Sick with the Visionary-T 3D snapshot sensor, the Portuguese vision integrator Neadvance with its specially developed 3D image processing solution for determining the position and orientation of crates and cartons was also involved in the implementation of robot-assisted depalletizing at a Czech car manufacturer.

Information on shape, alignment and spacing

This challenging task was solved with the Visionary-T 3D snapshot sensor. The streaming camera provides the robot controller with measured values on the shape, orientation and distance of objects on the pallet. The Visionary-T determines the distance and size of objects by measuring 3D time-of-flight (3DToF). To do this, the sensor emits infrared light, which is reflected by the objects in the vicinity. From the phase shift between light and light emission, the sensor calculates the distance to the pallet surface, i.e. to the individual cartons or crates, and displays this in a three-dimensional point cloud. Neadvance's algorithms use the three-dimensional image information to generate depth and intensity information for each individual pixel and thus determine the shape, level and orientation of boxes and cartons. This enables the robot controller to recognize which container can be approached next without collision and can be gripped best, and to calculate the corresponding target coordinates for the automatic path guidance of the robot.

In this application, the 3D snapshot sensor is mounted directly on the articulated robot. As the robot moves over the pallet, the sensor captures images of the pallet surface as it moves. The robust sensor design - durable metal housing, no moving parts inside - ensures a high level of reliability with low maintenance requirements. The high repetition rate of up to 50 images per second also ensures that the metrological precision and reliability of the robot guidance is maintained even during dynamic acceleration and reversing movements of the robot, as well as in the case of reflective surfaces or other difficult conditions. The robot approaches the pallet, picks up the goods safely with its special gripper and places the crate or carton on the conveyor belt. The process starts again from the beginning to pick up the next crate. The dynamics of the robot and the speed of the 3D detection of the pallet surface have considerably accelerated the depalletizing process, significantly increased the throughput in incoming goods and relieved employees of a repetitive task.

Joint engineering for the best performance

Sick's 3D snapshot technology and close collaboration with Neadvance during the engineering process resulted in a future-proof solution for robot-guided goods separation. The precise 3D measurement data and point clouds generated by the sensor made it possible to record loading scenarios more accurately and to further refine the complex gripping processes - even during commissioning and during operation. The new depalletizing solution works so reliably and efficiently that the car manufacturer is planning to use it in other plants as well.

Moritz Vogt, Applications Engineer 3D Compact Systems, and Darko Klinec, Product Manager 3D Compact Systems, Sick