Select XLD contours according to several features.

::select_contours_xld selects XLD contours from the input contours Contours according to the following features (parameter Feature):

- 'contour_length': all contours whose length is smaller than Min1
  or larger than Max1 are not returned (Min2 and Max2 have no
  influence here).

- 'maximum_extent': all contours whose maximum extent (as measured
  by their eight extremal points in row and column direction,
  according to Haralick und Shapiro: Computer and Robot Vision,
  Addison-Wesley 1992, chapter 3.2) is smaller than Min1 or larger
  than Max1 are not returned (Min2 and Max2 have no influence here).

- 'direction': only contours for which the direction of the
  regression line is between Min1 and Max1 (in radians,
  counter-clockwise) are returned (Min2 and Max2 have no influence
  here).

- 'curvature': only contours for which the mean distance to the
  regression line lies between Min1 and Max1, and for which the
  standard deviation of the distances is between Min2 and Max2 are
  returned.
  
- 'closed': only contours for which the distance between their 
  start point and their end point is less or equal
  Max2 pixels are returned.

Before contour can be filtered by ::select_contours_xld according to 'direction' or 'curvature', the parameters of the regression lines to the contours must be calculated by calling ::regress_contours_xld.

::regress_contours_xld

::get_contour_xld, ::get_contour_attrib_xld, ::gen_contours_skeleton_xld, ::lines_gauss, ::lines_facet, ::edges_sub_pix, ::get_regress_params_xld, ::get_contour_global_attrib_xld, ::query_contour_global_attribs_xld

R. Haralick, L. Shapiro: "Computer and Robot Vision" Vol. 1; Kapitel 3.2, Addison-Wesley 1992

Sub-pixel operators