Use statistical methods to predict whether an anomaly occurs in an image

Use a Statistical Prediction block to predict an anomaly by using statistical methods. That is, the block uses the output from a Measurement block to determine whether an anomaly occurs.

To use a statistical method to predict an anomaly

1. On the Block Definitions toolbar, select Prediction.

   The Prediction list expands.

2. Drag a Statistical Prediction block onto the Canvas.

   

3. Connect the output port from a Measurement block to the input port of the Statistical Prediction block.

4. Select the Statistical Prediction block on the Canvas.

5. Right-click the block, and then select Properties.

   

   The Properties side sheet appears on the right side of the screen.

   

6. In the Threshold value box, enter the absolute value for the threshold at which the anomaly occurs.

   If the absolute difference between the measured value and the mean measurement value is greater than the threshold, the anomaly is detected.

7. In the Deviation box, enter the standard deviation for the statistical model.

8. Select Apply Changes.

   The preview result value for the block updates on the Canvas. The preview value is the R value of the statistical prediction.