Skip to main content

Application example

Check the right type of fuel hoses and clips


The task

In automated automobile production, it is important to check whether the correct fuel hoses and clamps have been installed. The components are to be checked with the aid of a vision sensor.

The challenge

Due to the varying positions of the vehicles, the vision sensor must be able to recognize the components from different perspectives. Furthermore, the parts are limp, which makes position tracking with classic methods (based on contour or pattern comparison) more difficult.

Our solution

The VISOR® Object AI vision sensor reliably differentiates between the various hose and clamp types. By assigning a few example images to each class, the AI-based detector "Classification (AI)" automatically learns to distinguish between the different types. Variations in the position of the vehicle can be shown to the detector and it learns the necessary features.

Your advantages

  • Easy creation of example images for the different classes to train the detector
  • Much easier than setting up different detectors to determine rules
  • It is not necessary to set up a separate job for each class
  • The detector transmits the class type directly, which facilitates integration into the PLC

Recommended products for this application