4.2 Job in uniVision¶

Open the device website of the Machine Vision Device (by entering the IP address in the browser, by default 192.168.100.1), access the tab Jobs, and open the current job in the uniVision 3 software.

Load the template Pick objects with generic robot for easy setup.

Adjust the focus and brightness of the camera to get a sharp and well-illuminated image. By default, Trigger Mode is set to On and Trigger Source to Software (mandatory for picking). Set Trigger Mode to Off in order to adjust the camera image. Afterward, change the setting back to Software trigger.

Use Module Image Coordinate System in order to create a common frame between camera and robot. By default, the origin of the coordinate system is in the center of the camera image (if using a 5 MP camera). Use the same coordinate system at the robot (see 4.1 Three-Point Calibration at Robot). Use your markings on the picking plane to adjust the coordinate system.

Use Module Image Calibration to eliminate the lens distortion and to calculate the coordinates in mm. Buy the calibration plate ZVZJ (recommended) or print the corresponding PDF yourself on a flat and stiff material. Put the calibration plate in the field of view of the camera. Select the corresponding size of the calibration plate ZVZJ at Module Image Calibration, click on Capture Image and afterward on Calibrate. The Z Offset of -4 mm compensates for the height of the calibration plate ZVZJ. Adjust it if the height of your calibration plate is different.

Put your picking object below the camera and teach it in Module Image Locator (at sub-module Shape Models). Then, Module Image Locator finds the object according to the parameters.

Use Module Spreadsheet to create one string with the coordinates x, y and z and the rotations A, B and C. Enter the height of the picking object at Z. The following example works with robots based on XML communication (tested with KUKA robots). In general, the robot programmer and the user configuring the Machine Vision Device must agree on a common protocol.

Use Device TCP to send the result via socket messaging to the robot.

Save the job in the projects folder of the Machine Vision Device.