You are here: Start » Filter Groups » FitLineToPoints

FitLineToPoints

Description

Finds a line well approximating a set of points.

Filters in this group

Icon Name Description
FitLineToPoints Approximates points with a line using the Least Squares method.
FitLineToPoints_M Approximates points with a line using selected M-estimator for outlier suppression.

Finding a locally optimal line. Good enough when the number of outliers is small.

FitLineToPoints_RANSAC Approximates points with a line using a RANSAC algorithm.

Finds a well matching line, but for handling outliers requires a distance threshold that may be difficult to set.

FitLineToPoints_TheilSen Approximates points with a line using TheilSen algorithm, optionally with Siegel's improvement.

Finds a well matching line, ignoring up to 29.3% (TheilSen) or 50.0% (Siegel) outliers. Outliers do have some influence on accuracy.

FitLineToPoints_LTE Approximates points with a line using Least Trimmed Error algorithm.

Brute-force finding of a line that best matches a subset of the input points. Very efficient against outliers, but possibly slow for bigger subsets.