Recipe Editor

The Recipe Editor provides a simple workflow for users to change image capture settings, set up part alignment, lay out regions of interest (ROIs) for their inspection, collect and label model training data, configure process logic and IO behavior, and link the OV20i to an FTP server for data storage.

The recommended workflow for building a Recipe is:

Imaging Setup → Alignment Block → Region of Interest Block → Classification Block → IO Block

1. Navigation Bar - Click on the name of your Recipe in the Navigation Bar (i.e. “Assembly Check” in this example) to return to this page from any Recipe Editor sub-menu.

2. Recipe Name / PLC Recipe ID / Description - The Recipe information displayed here may be edited by clicking on the ✏️ Edit button to the right of each field.

3. Export Recipe - Click here to export a copy of the active Recipe to your computer. You can choose to export with or without library captures associated with the Recipe—training images will be included in all instances.

4. Imaging Setup - Adjust focus, exposure, gain, and lighting settings to obtain the best image quality for inspection, and select from different trigger options to suit your use case.

5. AI Blocks - Define your deep learning inspection task using intuitive configurators organized into 3 modules:

   • Alignment Block - Capture a Template Image as a reference for your inspection program, and use pattern-matching to locate and orient parts for relative inspection.

   • Region of Interest Block - Draw Regions of Interest (ROIs) on the Template Image to define key inspection areas, and group regions with similar appearances or failure modes into multiple Inspection Types.

   • Classification Block - Train a deep learning model to determine what’s inside a Region of Interest.

     OR

   • Segmentation Block - Train a deep learning model to take an image and identify each pixel's content.

6. IO Block - Define Pass/Fail rules for each captured image, and use a built-in Node-RED editor to customize digital IO functions, complex logic, and web-based dashboards and user interfaces.

7. FTP - Send captured images to a remote server using the File Transfer Protocol (FTP).