boofcv.alg.feature.shapes

Class SplitMergeLineFitLoop

java.lang.Object

boofcv.alg.feature.shapes.SplitMergeLineFit

boofcv.alg.feature.shapes.SplitMergeLineFitLoop

public class SplitMergeLineFitLoop
extends SplitMergeLineFit

Implementation of SplitMergeLineFit for looped lists of points. The beginning and end of the list are assumed to be connected. An additional check is done in the start to find two points which are far apart.

Field Summary

Fields

Modifier and Type	Field and Description
protected int	N

Fields inherited from class boofcv.alg.feature.shapes.SplitMergeLineFit

changed, contour, line, maxIterations, point2D, splits, toleranceMerge, toleranceSplitSq, work

Constructor Summary

Constructors

Constructor and Description
SplitMergeLineFitLoop(double toleranceSplit, double toleranceMerge, int maxIterations)

Method Summary

Methods

Modifier and Type	Method and Description
protected int	circularDistance(int start, int end) Distance the two points are apart in clockwise direction
protected boolean	mergeSegments() Merges lines together which have an acute angle less than the threshold.
void	process(java.util.List<georegression.struct.point.Point2D_I32> contour) Approximates the input list with a set of line segments
protected int	selectFarthest(java.util.List<georegression.struct.point.Point2D_I32> contour) Computes the distance between pairs of points which are separated by 1/2 the contour list.
protected int	selectSplitOffset(int indexStart, int length) Finds the point between indexStart and the end point which is the greater distance from the line (set up prior to calling).
protected void	splitPixels(int indexStart, int length) Recursively splits pixels between indexStart to indexStart+length.
protected boolean	splitSegments() Splits a line in two if there is a paint that is too far away

Methods inherited from class boofcv.alg.feature.shapes.SplitMergeLineFit

computeAcute, getSplits

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

N

protected int N

Constructor Detail

SplitMergeLineFitLoop

public SplitMergeLineFitLoop(double toleranceSplit,
                     double toleranceMerge,
                     int maxIterations)

See Also:

SplitMergeLineFit.SplitMergeLineFit(double, double, int)

Method Detail

process

public void process(java.util.List<georegression.struct.point.Point2D_I32> contour)

Description copied from class: SplitMergeLineFit

Approximates the input list with a set of line segments

Specified by:

process in class SplitMergeLineFit

Parameters:

contour - Ordered list of connected points.

splitPixels

protected void splitPixels(int indexStart,
               int length)

Recursively splits pixels between indexStart to indexStart+length. A split happens if there is a pixel more than the desired distance away from the two end points. Results are placed into 'splits'

selectFarthest

protected int selectFarthest(java.util.List<georegression.struct.point.Point2D_I32> contour)

Computes the distance between pairs of points which are separated by 1/2 the contour list. The index of the first pixel in the pair with the greatest distance is returned

Returns:

Index of the first pixel which should be used to split the list. The other end is ret + N/2

mergeSegments

protected boolean mergeSegments()

Merges lines together which have an acute angle less than the threshold.

Returns:

true the list being changed

splitSegments

protected boolean splitSegments()

Splits a line in two if there is a paint that is too far away

Returns:

true for change

selectSplitOffset

protected int selectSplitOffset(int indexStart,
                    int length)

Finds the point between indexStart and the end point which is the greater distance from the line (set up prior to calling). Returns the index if the distance is less than tolerance, otherwise -1

circularDistance

protected int circularDistance(int start,
                   int end)

Distance the two points are apart in clockwise direction