EMatcher::GetDontCareThreshold
EMatcher::SetDontCareThreshold
"Don't care" threshold.
Namespace: Euresys::Open_eVision
[C++]
OEV_UINT32 GetDontCareThreshold() const
void SetDontCareThreshold(OEV_UINT32 un32Threshold)
Remarks
If the pattern cannot be inscribed in a rectangle because there are foreign objects in a close neighborhood, mismatches can be avoided by using "don't care" pixels: all the pattern pixels whose value is strictly below DontCareThreshold
will be ignored.
By default, this property is set to 0
: no "don't care" pixel exists.
Note. When you use the "don't care" feature Geometrical property of a coded element., either the pattern is well contrasted from its background -then set DontCareThreshold
to an appropriate thresholding value- or it is not -then set the background pixels of the pattern to some low value (by a masking operation) and set DontCareThreshold
to this low value plus one.
In addition, EMatcher::AdvancedLearning
does not work when using a don't care threshold and a masked image.
However, it is compatible with the overload of EMatcher::LearnPattern
taking an ERegion
, which allows to do the same as EMatcher::DontCareThreshold
.
EMatcher.DontCareThreshold
"Don't care" threshold.
Namespace: Euresys.Open_eVision
[C#]
uint DontCareThreshold
{ get; set; }
Remarks
If the pattern cannot be inscribed in a rectangle because there are foreign objects in a close neighborhood, mismatches can be avoided by using "don't care" pixels: all the pattern pixels whose value is strictly below DontCareThreshold
will be ignored.
By default, this property is set to 0
: no "don't care" pixel exists.
Note. When you use the "don't care" feature, either the pattern is well contrasted from its background -then set DontCareThreshold
to an appropriate thresholding value- or it is not -then set the background pixels of the pattern to some low value (by a masking operation) and set DontCareThreshold
to this low value plus one.
In addition, EMatcher.AdvancedLearning
does not work when using a don't care threshold and a masked image.
However, it is compatible with the overload of EMatcher.LearnPattern
taking an ERegion
, which allows to do the same as EMatcher.DontCareThreshold
.
EMatcher.DontCareThreshold
"Don't care" threshold.
Module: open_evision
[Python]
DontCareThreshold: int
{get; set}
Remarks
If the pattern cannot be inscribed in a rectangle because there are foreign objects in a close neighborhood, mismatches can be avoided by using "don't care" pixels: all the pattern pixels whose value is strictly below DontCareThreshold
will be ignored.
By default, this property is set to 0
: no "don't care" pixel exists.
Note. When you use the "don't care" feature, either the pattern is well contrasted from its background -then set DontCareThreshold
to an appropriate thresholding value- or it is not -then set the background pixels of the pattern to some low value (by a masking operation) and set DontCareThreshold
to this low value plus one.
In addition, EMatcher.AdvancedLearning
does not work when using a don't care threshold and a masked image.
However, it is compatible with the overload of EMatcher.LearnPattern
taking an ERegion
, which allows to do the same as EMatcher.DontCareThreshold
.