Robot-Vision Calibration Using Multi-Camera Pose Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current machine vision systems for robots face challenges in accurately calibrating the relationship between the machine vision coordinate system and the robot coordinate system, leading to inaccuracies in workpiece pose measurement and task execution, especially when dealing with multiple cameras and directional precision issues.
Innovation Solution
A system and method that involves moving the robot to various poses while acquiring images of a calibration object with features at known positions, using machine vision to determine the relationship between the machine vision system's and robot's coordinate systems, and imposing constraints on the number of degrees of freedom to ensure accurate calibration of multiple cameras simultaneously.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine vision is used to measure workpiece pose, then measurement speed and non-contact sensing are improved, but coordinate system transformation accuracy deteriorates due to hand-eye calibration errors
Solution Approach 1:
A calibration object with known feature positions is introduced as an intermediary between the machine vision system and the robot. The calibration object serves as a common reference frame that enables accurate coordinate transformation by providing known geometric relationships that can be observed by the vision system and correlated with robot poses, thereby mediating the calibration process and improving transformation accuracy
Solution Approach 2:
The patent replaces traditional mechanical calibration methods with vision-based measurement. Instead of using physical measurement tools or contact-based sensors to establish coordinate relationships, the system uses machine vision to detect features on the calibration object and compute transformation matrices, substituting mechanical calibration procedures with optical measurement and computational geometry
2Adaptability or versatility
If multiple cameras are used for calibration, then measurement coverage and versatility are improved, but calibration complexity and computational burden increase
Solution Approach 1:
The patent merges the calibration processes of multiple cameras into a unified procedure. By using a single calibration object that can be observed by multiple cameras simultaneously, the system combines all camera calibrations into one coordinated process rather than performing separate calibrations for each camera, thereby reducing overall complexity while maintaining multi-camera capability
Solution Approach 2:
The calibration object is designed to serve multiple functions: it can be observed by multiple cameras simultaneously, it provides known geometric features for coordinate transformation, and it enables calibration of the entire robot-vision system. This universal calibration object handles multiple calibration tasks in a single setup, reducing the need for separate calibration procedures for each camera
3Ease of manufacture
If traditional hand-eye calibration techniques are used, then implementation simplicity is improved, but calibration accuracy deteriorates due to inconsistent pose estimates
Solution Approach 1:
The patent implements feedback by using the known feature positions on the calibration object to verify and refine the coordinate transformation. The system measures feature positions with the machine vision system, compares them against the known positions, and uses this feedback to compute accurate transformation matrices. This feedback mechanism ensures that the calibration accurately reflects the true relationship between coordinate systems
Solution Approach 2:
The patent changes the approach from estimating poses independently to directly computing transformation matrices based on measured feature positions and known geometry. By changing the calibration parameters from pose estimates to transformation matrices derived from precise feature measurements, the system achieves higher accuracy while maintaining implementation feasibility
Data Source
AI summary
A system and method for robustly calibrating a vision system and a robot is provided. The system and method enables a plurality of cameras to be calibrated into a robot base coordinate system to enable a machine vision/robot control system to accurately identify the location of objects of interest within robot base coordinates.