Configuring a workflow

You configure a workflow with the Station App. Configure the workflow according to your business needs. You have the following configuration choices.

• Workflow name and description — The name and optional description for the workflow.

• Image source — An image source is where the workflow gets images that it analyzes with a model. An image source can be a camera or a local folder on the edge device. If the image source is a folder, the process for adding images to the folder is determined by the needs of your business and is outside the scope this documentation.

• Model — The Amazon Lookout for Vision model that the workflow uses to analyze images.

• Manual workflow trigger — You can run a workflow manually by using the Station App. This lets your line operators choose when a workflow should run.

• Digital input trigger — A digital signal that triggers the running of a workflow. For example, you can specify that you want a GPIO.FALLING signal to start the workflow running. If you specify a digital input, you can still manually run the workflow.

• Outputs — Analysis results from the running of a workflow is stored on edge device for later use. You can get the location from the Station App. The analysis results include a predicted classification for the model and the model's confidence in the accuracy of its prediction. If the model is a segmentation model, you can see the masks and anomaly labels for any predicted anomalies.

  Optionally, you can configure a digital signal that a workflow sends when it completes analysis of an image with the model. You can also apply a rule that determines if the signal is sent. For example, you can specify that you want a Falling edge signal sent when the model predicts that an image contains an anomaly.

Before you can configure the workflow, make sure you have an image source and have deployed a model to the edge device. For more information, see Adding an image source and Deploying your model to a station.

To configure a workflow (Station App)

1. Open the Station App on your edge device by opening a browser and navigating to x.x.x.x:3000. For x.x.x.x, use the IP address of your edge device.

2. From the left menu, choose Management and then Workflows.

3. In the Workflows section, select the workflow that you want to configure.

4. Choose Edit workflow.

5. In Workflow details enter a name and optional description for the workflow.

6. In Image source and model, choose the image source (Image source) and model (Model) that you want to use in the workflow.

7. For Workflow trigger, choose one of the following:

   • If you want a line operator to manually trigger the running of the workflow in the Station App, choose Line operator or API call.

   • If you want a device pin signal to trigger the workflow, do the following:

     1. choose Digital input.

     2. For Signal type choose the type of signal that triggers the workflow (Rising edge or Falling edge).

     3. For Pin value, enter the input pin that sends the signal.

     4. For Debounce time enter the amount of time, in mulliseconds, that the workflow ignores repeated signals.

8. (Optional) If you want the workflow to send a digital signal after analyzing the image, do the following:

   1. In the Outputs section choose Add output.

   2. For Signal type choose the type of signal that you want the workflow to send. (Rising edge or Falling edge).

   3. For Pin value, enter the input pin that you want to workflow to send the signal to.

   4. For Pulse width enter the amount of time, in mulliseconds, that the workflow sends the signal for.

   5. For Rule, choose the rule that runs the workflow. You can choose to run the rule if the model predicts an anomalous image, a normal image, or all images, regardless of the model's prediction.

   6. Choose Save and to add the output.

      Note

      The digital signal can trigger the workflow as soon as you add the output.

9. Choose Save to save the workflow