Robot Camera Geometric Calibration for Accurate Coating Paths
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for calibrating the position of a camera on a multi-axis robot are inaccurate and complex, particularly in industrial settings where environmental disturbances and high-speed operations are common.
Innovation Solution
A method that adjusts the geometrical model of a multi-axis robot by determining the coefficients of transformation matrices using a systematic approach involving multiple image taking locations and remarkable points, without relying on the computation of mapped features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a geometrical model of the robot is used based on the assumption that the connection between the camera and the wrist is known and rigid and that the robot axes work perfectly, then the model is simple to implement, but the model does not correspond to physical reality due to mounting tolerances, camera position changes, and robot imperfections
Solution Approach 1:
The patent modifies the geometrical model by introducing correction parameters that account for mounting tolerances and robot imperfections. Instead of assuming perfect rigidity and perfect axes, the model incorporates adjustable parameters that are optimized to match actual physical measurements, thereby improving accuracy while maintaining model simplicity
Solution Approach 2:
The patent replaces the purely theoretical mechanical model with an empirical approach using camera images. By substituting the assumption of perfect mechanical behavior with actual visual measurements of the robot's position, the system compensates for mechanical imperfections without requiring complex mechanical corrections
2Measurement precision
If methods such as touching a spherical profile or rotating the camera around an axis are used to determine TCP, then measurement accuracy is improved, but the implementation becomes complex and time consuming
Solution Approach 1:
The patent enables the robot system to self-calibrate using its own camera and geometrical model. The system uses images captured by its own camera to determine its position and optimize model parameters, eliminating the need for external calibration equipment or complex manual procedures while maintaining high accuracy
Solution Approach 2:
The patent creates a virtual model (geometrical representation) of the robot and camera system that mirrors the physical system. By optimizing this virtual model to match actual camera images, the system achieves accurate TCP determination without requiring physical calibration artifacts or complex measurement procedures
3Device complexity
If calibration methods assume the robot is perfect, then the calibration computation is simpler, but the calibration is not reliable in industrial settings with environmental disturbances and high-speed operations
Solution Approach 1:
The patent performs calibration computations in advance to determine the geometrical model parameters and TCP position. By pre-optimizing the model before industrial operations begin, the system accounts for actual physical conditions without requiring complex real-time adjustments during high-speed operations or in disturbed environments
Solution Approach 2:
The patent replaces assumptions about perfect mechanical behavior with empirical data from camera images. By substituting theoretical mechanical perfection with actual visual measurements, the calibration becomes reliable for industrial conditions without requiring complex computations that account for every possible environmental variable
Data Source
AI summary
A geometrical model of a multi-axis robot equipped with a camera (30) comprises a first change matrix (TCamera→PG) and a second change matrix TPG→BF)). A method for adjusting the model comprises: aiming (1004), with the camera arranged in a first location (k1), at at least two points among L points of the target; determining the coordinates (Vk1,lCamera) of each point aimed at; moving (1005) the camera; d) aiming (1006), with the camera arranged in a second location (k2), at the same points (Pιι); determining the coordinates (Vk2,lCamera) of each point aimed at; calculating (1008), for each point aimed at, a difference of the difference (vk1,k2,l) between the coordinates nt in a base coordinate frame (BF), expressed in using the first and second change matrix (TCamera→PG, TkPG→BF); calculating (1010, 1018) an overall difference value (F, G). Variables (X1-X6, Δαi, Δli, Δθi, Δri) of coefficients of the first and second change matrices (TCamera→PG, TPG→BF) are determined by minimizing (1012, 1018; 1021) the overall value calculated during step g). The camera is successively brought into K locations. The product of the number (L) of points of the target by the number (K) of locations minus 6 (L*K−6) is greater than or equal to the number of variables of the geometrical model of the multi-axis robot.