Positioning and Assembly — Aluminum Ingot Marking
About 717 wordsAbout 2 min
1. Overview
This case explains how to use the planar target positioning and assembly (point cloud matching only) task to calculate the pose of an aluminum ingot from stable geometric features and output the target position and orientation for marking.
The workflow centers on point cloud registration for a single target. It does not depend on 2D instance segmentation or require training a visual model using the general target keypoint model workflow. For common operations such as camera connection, robot communication, and eye–hand calibration, this case lists the key checks and links to the existing guides for detailed instructions.
Software fields in this case
PickWiz still uses the UI labels Target, End-effector, and Pick Point. In this case, the Pick Point defines the positioning pose needed for marking; it does not mean that the ingot must be picked. The marking content, execution action, tool working distance, and safety interlocks must be configured for the actual equipment and process.
1.1 Configuration workflow
Confirm the features and imaging conditions
Create the task and set the ROI
Prepare the template and positioning point
Validate point cloud matching
Integrate and accept the process
1.2 Chapter navigation
- Scenario and preconfiguration checks: confirm the single-target conditions, feature stability, capture position, and positioning requirements.
- Create the task and basic configuration: select the point-cloud-only task, associate the equipment and calibration, and set the ROI.
- Point cloud template and marking position configuration: select distinctive structures, create the template, and configure the positioning point and tool orientation.
- Visual parameter configuration: check 3D preprocessing, coarse and fine matching, empty ROI detection, and positioning pose filters.
- Run and acceptance: validate historical samples, live capture, positioning without marking, and actual marking in sequence.
- Common requirements and troubleshooting: address repeated structures, reflections, positioning offsets, empty scenes, and cycle time.
1.3 Key configuration relationships
| Configuration object | Purpose | Required checks |
|---|---|---|
| Capture position | Keep the target features consistently visible | Features remain visible across the permitted incoming-position variation; target and camera are stable during capture |
| Point cloud template | Supply geometric features for single-target registration | Distinctive structure and stable relationship to the marking area |
| 3D ROI | Bound the area used for matching | Covers permitted feature movement without including several similar targets |
| Target Pick Point | Express the marking pose relative to the template | Position and orientation match the actual tool TCP and process |
| Eye–hand calibration | Transform the vision result into the coordinate relationship used by the equipment | Matches the current installation and robot pose at capture |
| Execution program | Mark using a validated positioning result | Checks result validity, abnormal responses, and safety interlocks |
Related documents
- Point cloud template creation guide: template capture, editing, coordinate adjustment, and downsampling.
- Point cloud matching pose estimation configuration guide: coarse and fine point cloud matching parameters.
- 3D preprocessing function configuration guide: ROI point cloud filtering, denoising, and downsampling.
- ROI configuration guide: workspaces and permitted positioning-point regions.
- Eye–hand calibration workflow: configure coordinate transformations for the installation.
- Visual computation configuration guide: data sources, storage, and output count.