Skip to content

minkowski_add1minkowski_add1MinkowskiAdd1MinkowskiAdd1minkowski_add1minkowski_add1🔗

Short description🔗

minkowski_add1minkowski_add1MinkowskiAdd1MinkowskiAdd1minkowski_add1minkowski_add1 — Perform a Minkowski addition on a region.

Signature🔗

minkowski_add1( region Region, region StructElement, out region RegionMinkAdd, integer Iterations )minkowski_add1( Region, StructElement : RegionMinkAdd : Iterations : )void MinkowskiAdd1( const HObject& Region, const HObject& StructElement, HObject* RegionMinkAdd, const HTuple& Iterations )static void HOperatorSet.MinkowskiAdd1( HObject region, HObject structElement, out HObject regionMinkAdd, HTuple iterations )def minkowski_add1( region: HObject, struct_element: HObject, iterations: int ) -> HObject

Herror minkowski_add1( const Hobject Region, const Hobject StructElement, Hobject* RegionMinkAdd, const Hlong Iterations )

Herror T_minkowski_add1( const Hobject Region, const Hobject StructElement, Hobject* RegionMinkAdd, const Htuple Iterations )

HRegion HRegion::MinkowskiAdd1( const HRegion& StructElement, Hlong Iterations ) const

HRegion HRegion.MinkowskiAdd1( HRegion structElement, int iterations )

Description🔗

minkowski_add1MinkowskiAdd1 dilates the input regions with a structuring element. By applying minkowski_add1MinkowskiAdd1 to a region, its boundary gets smoothed. In the process, the area of the region is enlarged. Furthermore, disconnected regions may be merged. Such regions, however, remain logically distinct region. The Minkowski addition is a set-theoretic region operation. It is based on translations and union operations.

Let \(M\) (StructElementstructElementstruct_element) and \(R\) (Regionregionregion) be two regions, where \(M\) is the structuring element and \(R\) is the region to be processed. Furthermore, let \(m\) be a point in \(M\). Then the displacement vector \({{\vec{v}}_{m}} = (dx,dy)\) is defined as the difference of the center of gravity of \(M\) and the vector \(\vec{m}\). Let \(t_{\vec{v}_{m}}(R)\) denote the translation of a region \(R\) by a vector \(\vec{v}\). Then

\[\begin{eqnarray*} \textrm{minkowski\_add1}(R,M) := \bigcup_{m \in M} {t_{\vec{v}_{m}} (R)} \end{eqnarray*}\]

For each point \(m\) in \(M\) a translation of the region \(R\) is performed. The union of all these translations is the Minkowski addition of \(R\) with \(M\). minkowski_add1MinkowskiAdd1 is similar to the operator dilation1Dilation1, the difference is that in dilation1Dilation1 the structuring element is mirrored at the origin. The position of StructElementstructElementstruct_element is meaningless, since the displacement vectors are determined with respect to the center of gravity of \(M\).

The parameter Iterationsiterationsiterations determines the number of iterations which are to be performed with the structuring element. The result of iteration \(n-1\) is used as input for iteration \(n\). From the above definition it follows that an empty region is generated in case of an empty structuring element.

Structuring elements (StructElementstructElementstruct_element) can be generated with operators such as gen_circleGenCircle, gen_rectangle1GenRectangle1, gen_rectangle2GenRectangle2, gen_ellipseGenEllipse, draw_regionDrawRegion, gen_region_polygonGenRegionPolygon, gen_region_pointsGenRegionPoints, etc.

Attention🔗

A Minkowski addition always results in enlarged regions. Closely spaced regions which may touch or overlap as a result of the dilation are still treated as two separate regions. If the desired behavior is to merge them into one region, the operator union1Union1 has to be called first.

Execution information🔗

Execution information
  • Multithreading type: reentrant (runs in parallel with non-exclusive operators).

  • Multithreading scope: global (may be called from any thread).

Parameters🔗

Regionregionregion (input_object) region(-array) → objectobjectHObjectHRegionHObjectHobject

Regions to be dilated.

StructElementstructElementstruct_element (input_object) region → objectobjectHObjectHRegionHObjectHobject

Structuring element.

RegionMinkAddregionMinkAddregion_mink_add (output_object) region(-array) → objectobjectHObjectHRegionHObjectHobject *

Dilated regions.

Iterationsiterationsiterations (input_control) integer → (integer)(integer)HTuple (Hlong)HTuple (int / long)intHtuple (Hlong)

Number of iterations.

Default: 11
Suggested values: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 15, 17, 20, 30, 40, 501, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 15, 17, 20, 30, 40, 50
Value range: 1 ≤ Iterations (lin)
Minimum increment: 1
Recommended increment: 1

Complexity🔗

Let \(F1\) be the area of the input region, and \(F2\) be the area of the structuring element. Then the runtime complexity for one region is:

\[\begin{eqnarray*} O(\sqrt{F1} \cdot \sqrt{F2} \cdot \textrm{Iterations}) \enspace . \end{eqnarray*}\]

Result🔗

minkowski_add1MinkowskiAdd1 returns 2 (H_MSG_TRUE) if all parameters are correct. The clipping according to the global image size is set via the operator set_system('clip_region',<'true'/'false'>). The behavior in case of empty or no input region can be set via:

  • no region: set_system('no_object_result',<RegionResult>)

  • empty region: set_system('empty_region_result',<RegionResult>)

Otherwise, an exception is raised.

Combinations with other operators🔗

Combinations

Possible predecessors

thresholdThreshold, regiongrowingRegiongrowing, connectionConnection, union1Union1, watershedsWatersheds, class_ndim_normClassNdimNorm, gen_circleGenCircle, gen_ellipseGenEllipse, gen_rectangle1GenRectangle1, gen_rectangle2GenRectangle2, draw_regionDrawRegion, gen_region_pointsGenRegionPoints, gen_region_polygon_filledGenRegionPolygonFilled

Possible successors

reduce_domainReduceDomain, select_shapeSelectShape, area_centerAreaCenter, connectionConnection

Alternatives

minkowski_add2MinkowskiAdd2, dilation1Dilation1

See also

transpose_regionTransposeRegion, minkowski_sub1MinkowskiSub1

Module🔗

Foundation