public static Cog3DModelFeatureGeneratorUsingCrsp2D3DsResults GenerateFeaturesModel3D(
	List<Cog3DCameraCalibration> raw2DFromPhys3Ds,
	List<List<Cog3DCrsp2D3D>> crsp2D3DsForAllPartPoses
)

public static Cog3DModelFeatureGeneratorUsingCrsp2D3DsResults GenerateFeaturesModel3D(
	List<Cog3DCameraCalibration> raw2DFromPhys3Ds,
	List<List<Cog3DCrsp2D3D>> crsp2D3DsForAllPartPoses
)

Public Shared Function GenerateFeaturesModel3D ( 
	raw2DFromPhys3Ds As List(Of Cog3DCameraCalibration),
	crsp2D3DsForAllPartPoses As List(Of List(Of Cog3DCrsp2D3D))
) As Cog3DModelFeatureGeneratorUsingCrsp2D3DsResults

Public Shared Function GenerateFeaturesModel3D ( 
	raw2DFromPhys3Ds As List(Of Cog3DCameraCalibration),
	crsp2D3DsForAllPartPoses As List(Of List(Of Cog3DCrsp2D3D))
) As Cog3DModelFeatureGeneratorUsingCrsp2D3DsResults

public:
static Cog3DModelFeatureGeneratorUsingCrsp2D3DsResults^ GenerateFeaturesModel3D(
	List<Cog3DCameraCalibration^>^ raw2DFromPhys3Ds, 
	List<List<Cog3DCrsp2D3D^>^>^ crsp2D3DsForAllPartPoses
)

public:
static Cog3DModelFeatureGeneratorUsingCrsp2D3DsResults^ GenerateFeaturesModel3D(
	List<Cog3DCameraCalibration^>^ raw2DFromPhys3Ds, 
	List<List<Cog3DCrsp2D3D^>^>^ crsp2D3DsForAllPartPoses
)

Return Value

A Cog3DModelFeatureGeneratorUsingCrsp2D3DsResults results object containing a List of the 3D model features and a List of the pose of the part in each part pose. The List of part poses in the results object is indexed by part pose index. The List of 3D model features is indexed by the FeatureModel3DIndex used in the specified crsp2D3DsForAllPartPoses. Thus, the crsp2D3Ds in crsp2D3DsForAllPartPoses with a FeatureModel3DIndex of N, were used in the generation of (Cog3DModelFeatureGeneratorUsingCrsp2D3DsResults).GetFeatureModel3DResults()[N].

If (Cog3DModelFeatureGeneratorUsingCrsp2D3DsResults).GetFeatureModel3DResults()[i].IsFound is false, it means that there was not enough information in the crsp2D3DsForAllPartPoses to generate 3D model feature[i].

The crsp2D3DsForAllPartPoses should be created by placing a single part under your cameras and simultaneously acquiring images of the one part from all the cameras. This set of images constitutes a "part pose". You should create several part poses of your part, see the Best Practices listed below for details. Providing multiple part poses improves the accuracy of the generated 3D model features.

After collecting the part poses (sets of images of the part), you need to extract the crsp2D3Ds from each of them. For each part pose, extract the crsp2D3Ds from the images of the part pose and put them into a List. The List of crsp2D3Ds from each part pose is then placed in another List thereby creating crsp2D3DsForAllPartPoses which is indexed as crsp2D3DsForAllPartPoses[partPoseIndex][crspIndex].

The 3D model features in the returned results object are in the Model3D coordinate space. The Model3D coordinate space is defined in part pose 0 or crsp2D3DsForAllPartPoses[0] i.e. the pose of the Phys3D origin in part pose 0 will be used to define the origin of the part's 3D coordinate space - Model3D. Thus, the pose of the part (the transform from Model3D space to Phys3D space) in crsp2D3DsForAllPartPoses[0] is the identity transform.

If crsp2D3DsForAllPartPoses[i][j].FeatureRaw2D is null or empty, then crsp2D3DsForAllPartPoses[i][j] is ignored during model feature generation.

All the specified camera calibrations must refer to the same Phys3D coordinate space.

Best Practices for Generating 3D Model Features:

• The FOV must contain a single part.
• Use the 5 part poses described below when creating crsp2D3DsForAllPartPoses.
  • Pose 0: Place the part as close as possible to the origin of Phys3D, use a z rotation of 0 degrees. Note that the origin of the Model3D coordinate space (the coordinate space all your 3D model features are defined in) is defined by the Phys3D origin in this pose.
  • Pose 1: Place the part in the upper left corner of FOV visible by both cameras, use a z rotation of ~20 degrees.
  • Pose 2: Place the part in the upper right corner of FOV visible by both cameras, use a z rotation of ~45 degrees.
  • Pose 3: Place the part in the lower left corner of FOV visible by both cameras, use a z rotation of ~67 degrees.
  • Pose 4: Place the part in the lower right corner of FOV visible by both cameras, use a z rotation of ~90 degrees.
• Verify the IsFound and the residuals in GetFeatureModel3DResults() from the returned results object are acceptable for your 3D model.