The cameras of the Pictor T-series by Vision & Control (Germany) have made it possible to implement the first pick-and-place jobs. The camera communicates with the machine's control unit in real time via a CANopen network or other networks. Its features include algorithms for determining the rotational orientation and position.
PICK-AND-PLACE IS A TYPICAL PROCESS when packaging and filling foods. From a machine vision perspective, a usual procedure is as follows: the system control unit triggers the image capture. The objects in the captured image are separated and identified, and their position and rotational orientation are determined. The results are then transmitted to the control unit and used to generate parameters for the drive system. Finally, the movement and picking process takes place.
In the example job, four Delta robots pick and fill 800 products per minute in the packaging department, i.e. each robot packs 200 parts per minute. As six parts are simultaneously picked during each picking cycle, this corresponds to approximately 33 picking cycles per minute or 1,8 seconds per picking cycle.
An intelligent camera is assigned to each robot. To achieve the desired performance with the Delta robots, the machine vision system must simultaneously determine the position and rotational orientation of at least six parts per picking cycle and transmit this data to the robots' control unit. This gives the machine vision system 300 milliseconds to determine the position and rotational orientation of each part.
News and reports