Hand-Eye Vision Calibration for Accurate Robot Pose Calculation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing robot systems face challenges in accurately calibrating the coordinate system for precise positioning and posture estimation of objects using vision sensors, particularly in hand-eye robot systems where the vision sensor is fixed on the robot's hand-tip part.

Innovation Solution

An arithmetic apparatus that includes a control part for outputting control signals to position an imaging part relative to a learning target object, generating a model through learning using imaging results, and performing arithmetic processing to calculate the position and posture of a processing target object based on learned parameters.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If coordinate system calibration is performed using traditional vision sensor methods, then the robot system can perform basic positioning tasks, but the positioning precision and posture estimation accuracy are insufficient for high-precision operations

Engineering Contradiction:
Improvepositioning precisionVSAvoidcoordinate system calibration accuracy
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The system performs preliminary coordinate system calibration and creates a parameter determination model before actual robot operations. By pre-calibrating the vision sensor coordinate system relative to the robot coordinate system and establishing mathematical models for position and posture calculation, the system ensures high measurement precision from the outset rather than attempting to achieve it during operation.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent replaces traditional mechanical positioning and measurement methods with vision-based measurement. The vision sensor captures images of calibration patterns and objects, and mathematical models convert these optical measurements into precise position and posture data, eliminating the need for complex mechanical measurement systems and improving both precision and reliability.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If the vision sensor is fixed on the robot hand-tip for hand-eye calibration, then the system can perform object manipulation tasks, but the coordinate system calibration becomes complex and accuracy is compromised

Engineering Contradiction:
Improvehand-eye operation capabilityVSAvoidcoordinate system calibration complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent introduces a calibration pattern as an intermediary object between the vision sensor and the robot coordinate system. This calibration pattern serves as a reference medium that facilitates the establishment of mathematical relationships between different coordinate systems, simplifying the calibration process while maintaining high accuracy for hand-eye operations.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system determines optimal parameters for position and posture calculation through the learned model, adjusting calculation parameters based on the specific geometric relationships and measurement conditions. This parameter optimization simplifies the calibration procedure while ensuring accurate coordinate transformation for versatile robot operations.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If traditional arithmetic processing is used without learned models, then the system operates with simpler processing, but the accuracy of position and posture calculation is insufficient

Engineering Contradiction:
Improveposture estimation accuracyVSAvoidarithmetic processing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs preliminary learning to determine optimal arithmetic processing parameters before actual position and posture calculation. By pre-processing calibration data and establishing mathematical models in advance, the system creates optimized calculation frameworks that deliver high measurement precision during operation without requiring complex real-time processing.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent creates mathematical models that copy and represent the geometric relationships observed during calibration. These models serve as simplified representations of complex spatial relationships, enabling accurate position and posture estimation through straightforward arithmetic operations rather than complex real-time calculations.

Inventive Principle:
Principle #26Copying

Data Source

PatentEP4596188A1Calculation device, calculation system, robot system, calculation method and computer program
Publication Date: 2025.08.06 NIKON CORP
  • EP4596188A1 patent drawingFigure 1~2
  • EP4596188A1 patent drawingFigure 3~4
  • EP4596188A1 patent drawingFigure 5

AI summary

An arithmetic apparatus includes: a control part that outputs a control signal for controlling an imaging part and a robot equipped with the imaging part; and a learning part that generates a model for determining a parameter of arithmetic processing in the control part, by learning using an imaging result of a learning target object by the imaging part. The control part outputs a first control signal. The learning part generates the model, by learning using learning image data generated by the imaging part imaging the learning target object in the predetermined positional relation, by control based on the first control signal. The control part performs the arithmetic processing, by using the parameter determined by the model and processing target image data generated by the imaging part imaging a processing target object, and calculates at least one of a position and posture of the processing target object.