Back to Adaptive Vision Library website

You are here: Start » Function Reference » Template Matching » LocateMultipleObjects_Edges


Finds all occurrences of a predefined template on an image by comparing object edges.



void avl::LocateMultipleObjects_Edges
	const avl::Image& inImage,
	atl::Optional<const avl::Region&> inSearchRegion,
	const avl::EdgeModel& inEdgeModel,
	int inMinPyramidLevel,
	atl::Optional<int> inMaxPyramidLevel,
	float inEdgeThreshold,
	bool inIgnoreEdgePolarity,
	bool inIgnoreBoundaryObjects,
	float inMinScore,
	float inMinDistance,
	atl::Array<avl::Object2D>& outObjects,
	atl::Optional<atl::Array<atl::Array<avl::Path>>&> outObjectEdges = atl::NIL,
	atl::Optional<int&> outPyramidHeight = atl::NIL,
	atl::Array<avl::Image>& diagEdgePyramid,
	atl::Array<avl::Image>& diagMatchPyramid,
	atl::Array<atl::Array<float> >& diagScores


Name Type Range Default Description
inImage const Image& Image on which object occurrences will be searched
inSearchRegion Optional<const Region&> NIL Region of possible object centers
inEdgeModel const EdgeModel& Model of objects to be searched
inMinPyramidLevel int 0 - 12 0 Defines the lowest pyramid level at which object position is still refined
inMaxPyramidLevel Optional<int> 0 - 12 3 Defines the total number of reduced resolution levels that can be used to speed up computations
inEdgeThreshold float 0.01 - 10.0f Minimum strength of edges used for matching with the model
inIgnoreEdgePolarity bool False Flag indicating whether edges of negated polarity should be ignored or not
inIgnoreBoundaryObjects bool False Flag indicating whether objects crossing image boundary should be ignored or not
inMinScore float 0.0 - 1.0 0.7f Minimum score of object candidates accepted at each pyramid level
inMinDistance float 0.0 - 10.0f Minimum distance between two found objects
outObjects Array<Object2D>& Found objects
outObjectEdges Optional<Array<Array<Path>>&> NIL Model edges of the found objects
outPyramidHeight Optional<int&> NIL Highest pyramid level used to speed up computations
diagEdgePyramid Array<Image>& Image edges used for matching at each pyramid level
diagMatchPyramid Array<Image>& Candidate object locations found at each pyramid level
diagScores Array<Array<float> >& Scores of the found objects at each pyramid level


The operation matches the object model, inEdgeModel, against the input image, inImage. The inSearchRegion region restricts the search area so that only in this region the centers of the objects can be presented. The inMinScore parameter determines the minimum score of the valid object occurrence. The inMinDistance parameter determines minimum distance between any two valid occurrences (if two occurrences lie closer than inMinDistance from each other, the one with greater score is considered to be valid).

In the inImage every pixel with gradient's magnitude at least inEdgeThreshold is considered an edge pixel. Only those are later used during the matching process. To establish the value of this threshold properly the diagEdgePyramid output which shows all the edges with sufficient gradient's magnitude may be used. If the inIgnoreEdgePolarity parameter is set, the object occurrences in the inImage image do not have necessarily to have the same contrast as the object in the model image.

The computation time of the filter depends on the size of the model, the sizes of inImage and inSearchRegion, but also on the value of inMinScore. This parameter is a score threshold. Based on its value some partial computation can be sufficient to reject some locations as valid object instances. Moreover, the pyramid of the images is used. Thus, only the highest pyramid level is searched exhaustively, and potential candidates are later validated at lower levels. The inMinPyramidLevel parameter determines the lowest pyramid level used to validate such candidates. Setting this parameter to a value greater than 0 may speed up the computation significantly, especially for higher resolution images. However, the accuracy of the found object occurrences can be reduced. Larger inMinScore generates less potential candidates on the highest level to verify on lower levels. It should be noted that some valid occurrences with score above this score threshold can be missed. On higher levels score can be slightly lower than on lower levels. Thus, some valid object occurrences which on the lowest level would be deemed to be valid object instances can be incorrectly missed on some higher level. The diagMatchPyramid output represents all potential candidates recognized on each pyramid level and can be helpful during the difficult process of the proper parameter setting.

The outObjects.Point array contains the model reference points of the matched object occurrences. The corresponding outObjects.Angle array contains the rotation angles of the objects. The corresponding outObjects.Match array provides information about both the position and the angle of each match combined into value of Rectangle2D type. Each element of the outObjects.Alignment array contains informations about the transform required for geometrical objects defined in the context of template image to be transformed into object in the context of corresponding outObjects.Match position. This array can be later used e.g. by 1D Edge Detection or Shape Fitting categories filters.


  • If an object is not detected, first try decreasing inMaxPyramidLevel, then try decreasing inMinScore.
  • If all the expected objects are correctly detected, try achieving higher performance by increasing inMaxPyramidLevel and inMinScore.
  • Please note, that due to the pyramid strategy used for speeding-up computations, some objects whose score would finally be above inMinScore may be not found. This may be surprising, but this is by design. The reason is that the minimum value is used at different pyramid levels and many objects exhibit lower score at higher pyramid levels. They get discarded before they can be evaluated at the lowest pyramid level. Decrease inMinScore experimentally until all objects are found.
  • If precision of matching is not very important, you can also gain some performance by increasing inMinPyramidLevel.
  • If the performance is still not satisfactory, go back to model definition and try reducing the range of rotations and scaling as well as the precision-related parameters: inAnglePrecision, inScalePrecision and inEdgeCompleteness.
  • To obtain high accuracy of matching make sure that the edges visible on the diagEdgePyramid output are not too thick. Increase inEdgeThreshold to make them thinner.
  • Also consider increasing inEdgeThreshold if there is much noise on the diagEdgePyramid output.
  • Verify accuracy of the matching by checking the outObjectEdges output. If the accuracy is lower than expected, make sure that inMinPyramidLevel = 0, try increasing inEdgeThreshold and also consider higher values for inAnglePrecision during model creation.
  • If the object being detected can be darker or brighter than the background on different images, use the inIgnoreEdgePolarity parameter set to True.
  • Define inSearchRegion to limit possible object locations and speed-up computations. Please note, that this is the region where the outObject.Point results belong to (and NOT the region within which the entire object has to be contained).


Locating multiple objects with the edge-based method.


For more information about local coordinate systems please refer to the following article.

Additional information about Template Matching can be found in Machine Vision Guide: Template Matching

Hardware Acceleration

This operation supports automatic parallelization for multicore and multiprocessor systems.

See Also

  • LocateSingleObject_Edges – Finds a single occurrence of a predefined template on an image by comparing object edges.