The Classification Block Logic node allows you to create rules that define overall pass/fail conditions for images in a recipe. This is used to define behavior of HMI, EthernetIP and Profinet. The output from this node is usually connected to one or both of the following…
Before configuring the Classification Block Logic, you must have completed setting up the ROI Block, and added all classes you plan to use in the Classification Block. This example is based on a simple recipe to detect whether a coin is heads or tails. The recipe has the following attributes:
To configure the Classification Block Logic node for this example:
<aside> 💡 For more complex recipes, this node provides a few other settings:
Add Rule: For recipes with multiple ROIs, you must add rules for each one of them using the button labeled “+add” in the lower left.
Set Confidence Threshold: You may optionally choose to add a lower limit for confidence for each rule if you want to augment how strict your inspection is. For example, maybe you want all ROIs to be “good” with >60% confidence to make sure that parts that are on the border of good or bad are rejected.
ALL/ANY rules must pass: If you add multiple rules to this node, you can use the drop-down menu at the bottom of the settings window to choose whether all or any of the rules must pass to count as a pass.
</aside>