Hand-Eye Calibration Using Runtime Workpiece Linear Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional machine vision hand-eye calibration methods are cumbersome and costly due to the need for special calibration targets and fixtures, especially in manufacturing processes where spatial constraints are a concern, leading to increased setup time and operational difficulties.
Innovation Solution
A system and method for performing 2D hand-eye calibration by using straight line features on a runtime work piece, eliminating the reliance on special calibration targets by establishing a coordinate system based on these features during motion, allowing calibration using a motion stage and vision system cameras to map pixel positions to physical positions in a motion coordinate system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional calibration targets and fixtures are used for hand-eye calibration, then calibration accuracy can be achieved, but device complexity and setup time increase significantly
Solution Approach 1:
The patent extracts the calibration function from specialized calibration targets and applies it to any runtime workpiece with recognizable features. By removing the dependency on dedicated calibration hardware, the system achieves calibration without complex fixtures while maintaining accuracy through feature-based recognition on ordinary workpieces.
Solution Approach 2:
The patent makes the workpiece itself serve dual purposes: both as the object being processed and as the calibration target. Any runtime workpiece with identifiable features can perform calibration, eliminating the need for separate calibration hardware and reducing overall system complexity.
2Measurement precision
If special calibration targets are used, then calibration can be performed, but setup time and operational difficulty increase
Solution Approach 1:
The system performs calibration automatically during the normal operation sequence without requiring preliminary setup of specialized targets. The calibration process is integrated into the workflow, using features already present on the workpiece when it enters the processing area, thereby eliminating dedicated setup time.
Solution Approach 2:
The workpiece itself provides the calibration references through its inherent features. The system uses recognizable patterns or markers already present on the workpiece to perform self-calibration, eliminating the need for external calibration targets and reducing setup operations to minimal automatic processing.
3Measurement precision
If custom calibration objects are used, then calibration accuracy is achieved, but manufacturing cost increases
Solution Approach 1:
The patent replaces expensive, custom-manufactured calibration targets with ordinary workpieces that are already part of the production process. These workpieces serve as disposable calibration references, eliminating the need to manufacture specialized calibration hardware while maintaining sufficient accuracy through feature-based recognition.
Solution Approach 2:
The system discards the need for separate calibration objects by utilizing the workpiece itself. The calibration function is recovered and integrated into the normal processing workflow, where the workpiece serves both as the object being manufactured and as the calibration reference, eliminating additional manufacturing costs.
Data Source
AI summary
This invention provides a system and method for hand-eye calibration of a vision system using an object under manufacture having at least one feature. The feature can be a linear feature and the object moves in at least one degree of freedom in translation or rotation on a motion stage. The system further comprises at least a first vision system camera and vision processor. The first vision system camera is arranged to acquire an image of the first linear feature on the object under manufacture and to track motion of the first linear feature in response to moving of the motion stage in at least one degree of translation. The first linear feature is identified in at least two different positions along a plane. The system computes a mapping between pixel positions in a pixel coordinate system of the first camera and physical positions in a motion coordinate system based upon locations of the at least two positions.