Point Cloud Matching Pose Estimation Configuration Guide
About 3033 wordsAbout 10 min
Point cloud matching pose estimation calculates a target’s pose through coarse and fine matching between a template point cloud and an instance point cloud. It is suitable for loading and unloading neatly arranged plane type targets, picking randomly arranged plane type targets, and positioning and assembly tasks.
Tuning order
First control the number of points in the template point cloud and verify coarse matching, then tune fine matching. Enable axis rotation pose adjustment only when a rotationally symmetric target has a consistent angular deviation about an axis.
Configuration overview
| Processing stage | Configuration items | Main purpose |
|---|---|---|
| Coarse matching preparation | Model Point Cloud Downsampling Size, Repeated Sampling Ratio Factor, Pose Angle Prior | Control the number of template points, sampling coverage, and known orientation constraints. |
| Contour coarse matching | Enable Contour Mode, Contour Mode, Coarse Registration Contour Search Radius | Use when the target has distinct contour features or keypoint coarse matching is unstable. |
| Fine matching | Object Pose Correction, Fine Matching Search Radius, Fine Matching Search Mode, Use Contour Mode, Contour Search Range | Further reduce the pose error between the template and instance point clouds. |
| Diagnostics and special corrections | Save Pose Estimation Data, Axis Rotation Pose Adjustment | Save diagnostic data or correct angular deviations about an axis for rotationally symmetric targets. |
Point cloud matching parameters
Related documents
Return to the overview of pose estimation methods
Clean the point cloud and extract contours before point cloud matching
Adjust pick points based on the matched pose