5.4 Detection Job in uniVision¶

Open the device website of the Machine Vision Device, select the tab Jobs, and open the current job in the uniVision 3 software.

Load the template Pick objects with 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. Adjust Trigger Mode to Off and switch to Run Mode in order to adjust the camera image. Afterward, change Trigger Source back to Software. If working with color images, make sure that Create BGRA Image is active at the input camera (only relevant if working with URCap).

Teach your object at Module Image Locator (or Module Image Pattern Match). For details, check the operating instructions of the software wenglor uniVision 3 (DNNF023).

In the template, Device Robot Vision is already pre-configured to send the results to the robot server.

Field	Explanation
Process Time [µs]	Process time to run the module in µs
Module State	Shows state of module: 0: No error Different to 0: Error
Output	Shows preview of the output sent to the wenglor robot server.
Shape Model Count	Defines the number of shape models (object types) used in the sub-module `Shape Model Height Difference`
Result Max Count	Defines the number of results used in the sub-module `Result List`.
Result True Count	Link the value with `Result True Count` of `Module Image Locator` or `Module Image Pattern Match` so that the robot knows how many objects are within the current image. By default, the value is 0. Linking `Result True Count` is mandatory so that the robot server knows how many objects are found in the current image.

Sub-Module Shape Model Height Difference (number of entries depends on parameter Shape Model Count):

Height Difference to Calibration [mm] #...x: Defines the height difference of each shape model to the calibration plate. By default, the value is 0.

Note

It is possible to pick different object types with different heights.
Enter the object height relative to the surface of the calibration plate in mm. If the object is lower than the calibration plate plane, enter a negative number. If the object is higher, enter a positive number. Also use the height difference if height variations occur after the calibration process.

Sub-Module Result List (number of entries depends on parameter Result Max Count)

Link the result list of Module Image Locator or Module Image Pattern Match to the Result List of Device Robot Vision.


Shape Model	Link the `Shape Model` (object type) for each result. By default, the value is 0.
X [px]	Link the x position for each result. Linking the x coordinate is mandatory.
Y [px]	Link the y position for each result. Linking the y coordinate is mandatory.
Phi (Z-Rotation)	[deg] Link the Phi value (z rotation) for each result. Linking Phi is mandatory.
Additional Value	Optionally link an additional string result (e.g. score value) for each result.

Note

It is possible to detect and pick multiple objects in a single image capture (if set up accordingly in the uniVision job).
Offsets in x and y can be set directly in Module Image Locator or Module Image Pattern Match.
Pixel results for x and y can only be used by the robot if they are distorted pixel coordinates. Modules without a calibration linked at the module calibration produce distorted pixel coordinates. If a calibration is linked at the module calibration, the module calculates undistorted pixel coordinates instead, which Device Robot Vision does not support. Make sure that only distorted pixel coordinates are linked at Device Robot Vision.

Save the detection job in the device projects folder on the Machine Vision Device. This job will later be used during the object detection in the robot program. Therefore, it is important to reuse the job name in the robot program.