The 2.5D vision system Eyes has been on the market since April and allows users to optically recognize parts with depth perception. Following the update, it now has a new landmark function, among other things. Landmarks are user-defined images that are attached to relevant objects such as workpieces or lifting devices. This enables the vision system to determine the exact position of the robot arm relative to them. This reduces the programming effort and enables the robot to grip more precisely. A landmark can also trigger the robot arm to perform certain actions.

Eyes enables fully automated processes
With the help of this new feature, users can easily combine their cobots with mobile robots. On the one hand, this works if the Eyes application is installed on a mobile robot that moves around the room. The landmark is then located at a fixed station, such as a production cell, a machine or a conveyor belt. When the robot approaches, it recognizes, thanks to Eyes, which job it should perform at the specific station and position and where it should go next. This allows users to send it on its own journey from station to station.

On the other hand, landmarks can initiate certain robot operations. If the Eyes application is installed stationary and a mobile robot with a landmark approaches, Eyes recognizes what the robot arm should do next. For example, if the transport robot is loaded with parts for order picking, the cobot starts to unload the workpieces.

"More and more manufacturers are looking for solutions to use their cobots more flexibly with the help of mobile transport robots," explains Enrico Krog Iversen, CEO of OnRobot. "This considerably extends the range of action of the robot arms and maximizes the benefits. With the help of the landmark function, Eyes makes interaction more efficient. It remains easy to use, so that companies of all sizes can continue to benefit."

New features for automated quality inspection
In addition to landmark calibration, the update also improves Eyes' ability to detect and grip randomly arranged objects. With the newly integrated color and blob detection features, the system can identify workpieces based on their color and shape alone. The user only has to feed in the information on the corresponding properties, while teaching is no longer necessary. This saves time during programming. In addition, Eyes now identifies different objects simultaneously with a single image. This allows the robot arm to act faster, which increases throughput. In addition, Eyes now allows users to request specific workpiece types and set precise parameters for the gripper distance. This ensures that the cobot also recognizes objects that are close together and does not damage them when gripping.

With the help of improved part recognition, Eyes can check workpieces for production errors after CNC machining or injection molding, for example. It is also conceivable that the system could be used for manual quality inspection. as