3D Coarse Matching V2

Function

Coarsely match the point cloud model with the original point clouds and output the coarsely calculated candidate poses of the target objects.

Sample Scenario

Generally used to coarsely find the target objects in the point clouds and obtain their poses. This Step is usually followed by the 3D Fine Matching Step.

Input and Output

../../../../_images/3d_coarse_matching_v2_input_and_output.png

Parameters

Parameters Tuning Level

Two tuning levels are provided, Basic and Advanced. All parameters in the Basic level are also available in the Advanced level.

Basic Tuning Level

Model Settings

Model Selection
Instructions:

  • Click on ▼ to select a point cloud model saved in the model gallery (the \resource\3d_matching directory under the project folder).

  • Once you select a point cloud model, its path and the path of its geometric center file are automatically displayed in Model File and Geo Center Point File below.

Note

The drop-down list of ▼ only displays the folders containing point cloud models and the associated geometric center/pick point files in the model gallery.

You can use Matching Model and Pick Point Editor to easily generate, edit and save point cloud models and the corresponding geometric centers/pick points.

Model File

Display the path of the selected point cloud model file.

Instructions: click on icon_file to select a point cloud model file in PLY format.

Geo Center Point File

Display the path of the selected geometric center file.

Instructions: click on icon_file to select a geometric center file in JSON format.

Note

If the point cloud model file and geometric center file you selected are not located in the same folder in the model gallery, the display of Model Selection changes to Custom.

../../../../_images/custom_model1.png
Matching Mode
Options: Surface Matching and Edge Matching
Instructions: please select the option that corresponds to the point cloud model you are using.

  • If the selected point cloud model represents the object edge, as shown below, please select Edge Matching.

    ../../../../_images/3d_coarse_matching_v2_edge.png

  • If the selected point cloud model represents the object surface, as shown below, please select Surface Matching.

    ../../../../_images/3d_coarse_matching_v2_face.png

Preprocessing Settings

Expected Point Count of Sampled Model
Set the expected point count of the point cloud model after sampling. The number of points in the point cloud model will approach this value after sampling.
Default value: 300
Instructions: the lower this value, the fewer points in the sampled point cloud, and the lower the matching accuracy is.

Set value

300

600

Point cloud model

icon_limit1

icon_limit3

Point count

276

632
Point Count Upper Threshold of Sampled Input Point Cloud
Set the upper limit of number of points in the input point cloud after sampling. If the number of points in the input point cloud after sampling exceeds this value, this point cloud is ignored and empty data will be output.
Default value: 50000

If the number of points in the input point cloud after sampling exceeds this value, this point cloud is ignored and empty data will be output, and the following error message is displayed.

../../../../_images/3d_coarse_matching_v2_limit.png

Pose Verification Settings

Expected Number of Detected Poses per Input Point Cloud
Set the expected number of matched poses for each input point cloud.
Default value: 3

Set value

1

3

Output

icon1

icon2

Advanced Tuning Level

Voting Settings

Upper Limit of Point Pair Count per Feature
The upper limit of the number of point pairs contained in each feature during the model analysis.
Default value: 50
Instructions: the lower this value, the faster the execution, but the accuracy is also lower.
Distance Quantification
The value for the quantification of the distance between points (distance interval = distance quantification × sampling interval).
Default value: 1
Instructions: the greater this value, the less accurate the result tends to be.
Angle Quantification
The quantification of the angle between points’ normals (angle interval = 360° / angle quantification)
Default value: 60
Instructions: the greater this value, the smaller the angle interval, and therefore the higher the accuracy of the matching result is, but a higher quality of point cloud is required.
Vote Ratio Lower Threshold
Poses with scores higher than the highest score multiplied by Vote Ratio Lower Threshold in voting will go through pose verification.
Default value: 0.8
Instructions: the lower this value, the more poses are used for pose verification, and therefore the more likely an exact match will be found, but the Step will also take longer to execute.
Referring Point Sampling Step
This parameter is used to sample points in the point cloud to make point pairs.
Default value: 5
Instructions: the greater this value, the faster the execution, but the accuracy is also lower.
Referred Point Sampling Step
This parameter is used to sample points in the point cloud to make point pairs.
Default value: 1
Instructions: the greater this value, the faster the execution, but the accuracy is also lower.
Referring points are sampling points on the point cloud model. Referred points are sampling points not on the point cloud model.
Referring points and referred points form point pairs. The larger the sampling steps, the fewer the point pairs, and therefore the faster the execution is.

Pose Filtering Settings

Use Distance NMS
Set whether to filter out candidate poses whose distances to the selected poses are less than 0.1 times the diameter of the object.
Default setting: on
../../../../_images/3d_coarse_matching_v2_posefilter.png
Filter Candite Pose by Axis
The pose with a significant angle difference between the constraint axis and the reference direction (larger than the angle difference upper threshold) will not be considered pose candidates. This parameter is usually enabled to filter out poses corresponding to mirror matching results.
Default setting: off

Pose Verification Settings

Voxel Length Generation Strategy
Set the generation strategy of the voxel length.
Default value: Auto
Instructions: for new users, Auto is recommended. If you select Manual, the Voxel Length field will appear below for you to set the value manually.

Note

If you divide the space occupied by the point cloud into a 3D grid, each unit cube in this grid is a voxel. Voxel length is the length of the cube edge.

Voxel Length (Displayed if Voxel Length Generation Strategy is set to Manual)
Set the length of the voxel.
Default value: 0.003
Instructions: the greater this value, the faster the execution, but the result also tends to be less accurate.
Voxel Length Lower Limit
Default value: 0.001
Voxel Length Upper Limit
Default value: 0.015

Result Verification

Visualization Option
Options: Sampled Model, Sampled Scene, Registration Result, and Points Used in Pose Verification
The following table provides example visualized output of each option.

Option

Example

Sampled Model The point cloud model after sampling

icon3

Sampled Scene The input point cloud after sampling

icon4

Registration Result The matching results

icon5

Points Used in Pose Verification The points used in pose verification

icon6