3D Vision Pose Measurement with Iterative Imaging Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional position and posture measurement systems for robots require lengthy teaching and measurement times, leading to inaccuracies in three-dimensional position and posture measurement, which is not suitable for applications needing precise motion control.
Innovation Solution
A position and posture measurement system that includes a calibrated visual sensor, a position and posture moving unit, calibration data storage, image processing, and correction units to iteratively adjust the imaging position and process images, allowing for accurate measurement and correction of the robot's position and posture relative to the workspace.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If conventional visual sensor measurement methods are used to detect three-dimensional positions of multiple reference points, then measurement coverage is improved, but measurement time and teaching time increase significantly
Solution Approach 1:
The patent extracts and focuses on measuring only the essential reference points (three non-collinear reference points) rather than attempting to measure all possible reference points in the workspace. This selective extraction of critical measurement targets reduces the number of teaching points required while maintaining sufficient measurement coverage for robot position and posture correction.
Solution Approach 2:
The patent segments the measurement process into two distinct phases: a calibration phase where the visual sensor's optical center position is determined, and a measurement phase where only three specific reference points are measured. This segmentation allows the system to prepare measurement parameters in advance, reducing real-time measurement time while maintaining accuracy.
2Measurement precision
If multiple measurement targets at multiple locations are imaged to ensure comprehensive coverage, then measurement accuracy is improved, but the number of teaching points increases and measurement time extends
Solution Approach 1:
The patent changes the measurement parameters by using a calibrated visual sensor with a determined optical center position. By pre-calibrating the sensor and storing the optical center information, the system transforms the measurement process from requiring multiple teaching points to requiring only three reference points, thereby reducing complexity while maintaining precision.
Solution Approach 2:
The patent performs preliminary calibration of the visual sensor before actual measurement, determining and storing the optical center position in advance. This preliminary action eliminates the need for repeated teaching of sensor parameters during measurement, reducing the number of teaching points required while ensuring measurement accuracy.
3Stability of the object's composition
If the robot repeats the same motion to perform operations, then operational consistency is maintained, but position and posture deviations accumulate making operations inaccurate
Solution Approach 1:
The patent implements a feedback mechanism where the visual sensor measures the actual position and posture of the robot relative to the workspace using three reference points. The measured deviations are fed back to correct the robot's motion parameters, ensuring that each operation accounts for accumulated deviations and maintains accuracy despite repeated motions.
Data Source
AI summary
Provided is a system that is capable of accurately measuring the three-dimensional position and posture of an object under measurement with simple teaching and in a short measurement time. A position and posture measurement system according to the present invention is provided with: a vision sensor that has already been calibrated; a position and posture moving unit that is capable of moving the three-dimensional position and posture of an object under measurement or the vision sensor; a calibration-data storage unit; an image processing unit that processes a captured image of the object under measurement; a position and posture measurement unit that measures the three-dimensional position and posture of the object under measurement by using the result of the image processing; an image-capturing-position correction unit that executes correction processing at least once, the correction processing being processing in which the vision sensor or the object under measurement is moved by the position and posture moving unit by using the measured three-dimensional position and posture and in which the image processing and the measurement of the three-dimensional position and posture are then sequentially carried out; and an image-capturing-position correction termination unit that executes the correction processing again upon determining that the correction processing is not to be terminated, while adopting the three-dimensional position and posture calculated either the last or in the middle as a three-dimensional position and posture to be used for robot control upon determining that the correction processing is to be terminated.