Operators

union_straight_contours_xld (Operator)

Name

union_straight_contours_xld — Compute the union of neighboring straight contours that have a similar direction.

Signature

union_straight_contours_xld(Contours : UnionContours : MaxDist, MaxDiff, Percent, Mode, Iterations : )

Description

union_straight_contours_xld merges neighboring XLD contours Contours if certain criteria are fulfilled. At each iteration, at most two contours that fulfill the given criteria are merged. The parameter Iterations controls how often this step is executed.

Two contours are merged if the shortest distance between their end points (end points are the projections of the contours' first and last points onto its regression line) is smaller than MaxDist, and if the difference in direction (i.e., the regression lines' direction) is smaller than MaxDiff (radians).

If only one of the criteria is fulfilled, the decision on merging can be influenced by the weighting factor Percent, which allows the exceeding of one limit to be balanced by the other value remaining below the limit by the same amount. The end point distance is weighted by Percent, while the directional difference is weighted by 100 - Percent.

This means that two contours are merged if they fulfill the following condition:

(shortest distance)/MaxDist*Percent + (direction difference)/MaxDiff*(100-Percent) <= 100

If, for example, two contours have an end point distance of 5.0 and a directional difference of 0.5 (with the limits chosen being MaxDist = 4.0 and MaxDiff = 0.625), both values differ from the limits by 25%. By choosing Percent = 60%, the larger end point distance is weigthed more than the smaller directional difference, and thus the contours are not merged. Contrary, if Percent = 40% is chosen, the contours are merged.

For Percent = 100%, only the end point distance is taken into account, while for Percent = 0% only the difference of direction is used. If Percent = 50% both criteria are equally valid.

In case there are parallel contours, there is a danger of merging neighboring contours. If this is to be avoided, Mode = 'noparallel' has to be chosen, while otherwise Mode = 'paralleltoo' suffices. For Mode = 'every', contours are merged unconditionally. All other parameters have no influence in this case.

The parameters of the regression line are calculated anew for merged contours.

Attention

Before the contour parameters can be returned by get_regress_params_xld, the parameters of the regression line to the contour must be calculated by calling regress_contours_xld.

Parallelization

• Multithreading type: reentrant (runs in parallel with non-exclusive operators).
• Processed without parallelization.

Parameters

Contours (input_object)  xld_cont-array object

Input XLD contours.

UnionContours (output_object)  xld_cont-array object

Output XLD contours.

MaxDist (input_control)  real (real)

Maximum distance of the contours' endpoints.

Default value: 5.0

Maximum difference in direction.

Default value: 0.5

Percent (input_control)  real (real)

Weighting factor for the two selection criteria.

Default value: 50.0

Mode (input_control)  string (string)

Should parallel contours be taken into account?

Default value: 'noparallel'

List of values: 'every', 'noparallel', 'paralleltoo'

Iterations (input_control)  string (string / integer)

Number of iterations or 'maximum'.

Default value: 'maximum'

Suggested values: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 'maximum'

Typical range of values: 1 ≤ Iterations ≤ 500 (lin)

Minimum increment: 1

Recommended increment: 1