3D Coarse Matching (Legacy)

Function: 3D point cloud coarse matching, use ppf feature point pair to match the model point cloud in the input scene point cloud.

Input Parameters:

Name	Type	Valid Range	Default Value	Description
Scene point cloud	NormalPoints	None	None	Scene point cloud

Name

Type

Valid Range

Default Value

Description

Scene point cloud

NormalPoints

None

Scene point cloud

Output Parameters:

Name	Type	Valid Range	Default Value	Description
Matching results	PosesList	None	[]	Matched results

Name

Type

Valid Range

Default Value

Description

Matching results

PosesList

None

[]

Matched results

Parameter Settings:

Name	Type	Valid Range	Default Value	Description
Model point cloud file	File	['.ply']	None	Model with normal point cloud, input m * n * 6 point cloud data path
Distance discrete quantity	Integer	[1, 500]	20	The number of distance discretization, the number of diameter discretization of the Model, can be calculated according to this parameter distanceDiscreteStep = diameter/distanceDiscreteNum
Angular discrete quantity	Integer	[1, 500]	30	The number of angles discrete, according to which angleDiscreteStep = 2 * PI/distanceDiscreteNum
Reference point step size	Integer	[1, 50]	5	Select the step size of the reference point, that is, select one reference point per referencePointStep points. Recommended value: 5
Deductive coefficient	Float	[0.0, 1.0]	0.5	Greater than 0 is a valid value, less than 0 will not be deduplicated, equal to 1.0 will remove all overlapping results, greater than 1 will reduce the number of successful matches
Clustering angle threshold (unit: degree)	Float	[0.0, 20.0]	15	When the transformation matrix is clustered, it is judged whether the two matrices belong to the same type of angle threshold. The larger the value, the smaller the number of clusters, and the more inaccurate the result may be. Recommended: 5, 10, 15
Clustering distance threshold (unit: mm)	Float	[0.0, 20.0]	15	When the transformation matrix is clustered, it is judged whether the two matrices belong to the same class of distance threshold. The larger the value, the smaller the number of clusters, and the more inaccurate the result may be. Recommended: 5, 15, 25
Voting filter threshold	Float	[0.0, 1.0]	0.1	When clustering, the voting filtering threshold relative to the maximum number of votes, the pose less than (maximum number of votes * threshold) will be filtered out. The larger the value, the fewer results may be matched. The recommended parameter is 0.05. When the cluster time is too long, this parameter can be appropriately adjusted.
Minimum number of votes	Integer	[0, 100000]	15	The minimum number of votes scored. For the pose voted out, the number of votes below that number will be directly filtered out, usually 2 to 15
Upper limit of the number of output pose	Integer	[0, 100000]	10000	The upper limit of the number of output pose, when the input is multiple point clouds, this upper limit refers to the upper limit of the matching result of each point cloud.

Name

Type

Valid Range

Default Value

Description

Model point cloud file

File

['.ply']

None

Model with normal point cloud, input m * n * 6 point cloud data path

Distance discrete quantity

Integer

[1, 500]

The number of distance discretization, the number of diameter discretization of the Model, can be calculated according to this parameter distanceDiscreteStep = diameter/distanceDiscreteNum

Angular discrete quantity

Integer

[1, 500]

The number of angles discrete, according to which angleDiscreteStep = 2 * PI/distanceDiscreteNum

Reference point step size

Integer

[1, 50]

Select the step size of the reference point, that is, select one reference point per referencePointStep points. Recommended value: 5

Deductive coefficient

Float

[0.0, 1.0]

0.5

Greater than 0 is a valid value, less than 0 will not be deduplicated, equal to 1.0 will remove all overlapping results, greater than 1 will reduce the number of successful matches

Clustering angle threshold (unit: degree)

Float

[0.0, 20.0]

When the transformation matrix is clustered, it is judged whether the two matrices belong to the same type of angle threshold. The larger the value, the smaller the number of clusters, and the more inaccurate the result may be. Recommended: 5, 10, 15

Clustering distance threshold (unit: mm)

Float

[0.0, 20.0]

When the transformation matrix is clustered, it is judged whether the two matrices belong to the same class of distance threshold. The larger the value, the smaller the number of clusters, and the more inaccurate the result may be. Recommended: 5, 15, 25

Voting filter threshold

Float

[0.0, 1.0]

0.1

When clustering, the voting filtering threshold relative to the maximum number of votes, the pose less than (maximum number of votes * threshold) will be filtered out. The larger the value, the fewer results may be matched. The recommended parameter is 0.05. When the cluster time is too long, this parameter can be appropriately adjusted.

Minimum number of votes

Integer

[0, 100000]

The minimum number of votes scored. For the pose voted out, the number of votes below that number will be directly filtered out, usually 2 to 15

Upper limit of the number of output pose

Integer

[0, 100000]

10000

The upper limit of the number of output pose, when the input is multiple point clouds, this upper limit refers to the upper limit of the matching result of each point cloud.