Cog3DPartCorresponderUsing2DPoses.Execute Method

Cog3DPartCorresponderUsing2DPoses Execute Method

Cognex VisionPro 9.7

Method to run the corresponder based on the 2D part poses for each part instance in each camera.

Namespace: Cognex.VisionPro3D
Assembly: Cognex.VisionPro3D (in Cognex.VisionPro3D.dll) Version: 73.0.0.0

Syntax

public List<Dictionary<int, int>> Execute(
	List<List<CogTransform2DLinear>> raw2DFromModel2Ds
)

public List<Dictionary<int, int>> Execute(
	List<List<CogTransform2DLinear>> raw2DFromModel2Ds
)

Public Function Execute ( 
	raw2DFromModel2Ds As List(Of List(Of CogTransform2DLinear))
) As List(Of Dictionary(Of Integer, Integer))

Public Function Execute ( 
	raw2DFromModel2Ds As List(Of List(Of CogTransform2DLinear))
) As List(Of Dictionary(Of Integer, Integer))

public:
List<Dictionary<int, int>^>^ Execute(
	List<List<CogTransform2DLinear^>^>^ raw2DFromModel2Ds
)

public:
List<Dictionary<int, int>^>^ Execute(
	List<List<CogTransform2DLinear^>^>^ raw2DFromModel2Ds
)

Parameters

raw2DFromModel2Ds: Type: System.Collections.Generic List List CogTransform2DLinear
A doubly indexed list of 2D part poses found in each camera, indexed first by camera, then by part instance. May not be null.
Note that the indices of raw2DFromModel2Ds specify the camera indices, and original part instance indices.

Return Value

Type: List Dictionary Int32, Int32
The part correspondences among different cameras. Each item is a Dictionary for one 3D part instance, and each KeyValuePair inside the Dictionary corresponds to (cameraIndex, originalPartInstanceIndex). Note that originalPartInstanceIndex is the index of the corresponding item in raw2DFromModel2Ds.

In other words, if raw2DFromModel2Ds[i][j] belongs to one 3D part instance, then the Dictionary for that 3D part instance has KeyValuePair of (i, j).

Exceptions

Exception	Condition
ArgumentNullException	If raw2DFromModel2Ds is null.
ArgumentException	If raw2DFromModel2Ds.Count is not the same as the number of cameras specified by the input parameter raw2DFromPhys3Ds in the constructor.

Remarks

Usage model is as follows:

Run pattern finding tools ( such as trained PatMax) to find the 2D poses of part instances in each camera (This step generates raw2DFromModel2Ds).
Run this method to compute part correspondences among different cameras.
Run feature extraction for each corresponded part.
If there are enough features extracted for one part instance, run pose estimation for the corresponded part.