Constructing a Best Fit or Best Fit Recompensate Line

Best Fit Recompensate is only accurate when you construct a line using points.

You can construct a "best fit" two or three-dimensional line from two or more features. The best fit construction method takes the actual measured points while the best fit recompensate method takes the ball's center. In both cases, the average squared error is minimized in the least squares method and the maximal error is minimized in the min/max method. You can also choose to remove outliers or apply a Gaussian filter to the constructed line.

To construct a best fit or best fit recompensate line:

  1. Open the Construct Line dialog box (Insert | Feature | Constructed | Line).

  2. From the Method list, select either the Best Fit or BF Recomp option.

  3. From the Feature list, select at least two features.

  4. Select either the 2D or 3D option.

For Best Fit Recompensate, one feature must be a point.

  1. If desired, click the Remove outliers check box and specify a value in the Standard deviation multiple box.

  2. If desired, click the Apply gauss filter check box and specify a value in the Cuttoff wavelength box.

  3. If you want to change the feature theoretical values, select the Feature theoreticals check box and type in the values. For details, see the "Specifying Feature Theoreticals" topic in the PC-DMIS Core documentation.

  4. Click the Create button.

The Edit window command line for the Best Fit method would read:

CONSTR/LINE,BF, feat_1,feat_2, …

OUTLIER_REMOVAL/(ON | OFF), stdDevMultiple

FILTER/(ON | OFF),WAVELENGTH=ctfoffWavelength

(This method uses the actual measured points for construction.)

The Edit window command line for the Best Fit Recompensate method would read:

CONSTR/LINE,BFRE,feat_1,feat_2, …

OUTLIER_REMOVAL/(ON | OFF), stdDevMultiple

FILTER/(ON | OFF),WAVELENGTH=cutoffWavelength

(This method uses the center of the probe for the measurement, and then recompensating after measuring the features.)

Constructing a Line from Two or More Features

More:

Remove Outliers / Standard Deviation Multiple

Apply Gauss Filter / Cutoff Frequency