Robot Camera Pose Estimation for Offline Teaching Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robot control systems struggle to accurately operate robots when there are deviations in the relative positional relationship between the robot and the camera, and they are not applicable when the camera cannot be mounted on the robot, leading to inaccuracies in simulating operations.
Innovation Solution
A control system comprising a simulation device, a robot controller, an imaging device, and an estimation module that estimates the position and pose of the imaging device relative to the robot, generates simulation images, and determines correction amounts to align the robot's actual position and pose with the simulated model, using offline teaching and image processing to minimize operational errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the camera is mounted on the robot with a fixed relative positional relationship, then the robot can be operated according to simulation, but the system becomes inapplicable when the camera cannot be mounted on the robot or when deviations occur in the relative positional relationship
Solution Approach 1:
The patent creates a virtual camera in the simulation environment that replicates the imaging characteristics of the actual camera. By copying the camera's position, pose, and imaging parameters into the virtual space, the system can accurately simulate robot operations regardless of whether the physical camera is mounted on the robot or elsewhere, resolving the adaptability issue while maintaining simulation accuracy
Solution Approach 2:
The patent introduces an imaging device position/pose estimation module as an intermediary that bridges the actual camera and the virtual camera. This module estimates the camera's position and pose relative to the robot based on captured images, and transmits this information to the simulation device to configure the virtual camera accordingly, enabling accurate simulation without requiring fixed physical mounting
2Measurement precision
If calibration is performed to align simulation images with actual images, then operational accuracy is improved, but the calibration process requires significant time and manual intervention
Solution Approach 1:
The patent implements automatic calibration where the system itself performs the alignment task. The imaging device captures actual robot images, the estimation module automatically determines camera position and pose, and the simulation device automatically adjusts the virtual camera parameters to match. This self-calibrating process eliminates manual intervention and significantly reduces calibration time while achieving high alignment accuracy
Solution Approach 2:
The system uses feedback from the actual captured images to continuously refine the virtual camera parameters. By comparing actual robot images with simulation images and using the deviation information to adjust the virtual camera's position and pose, the system achieves accurate alignment automatically, reducing both manual calibration time and improving precision
Data Source
AI summary
A control device estimates a position and pose of an imaging device relative to a robot based on an image of the robot captured by the imaging device. A simulation device arranges a robot model at a teaching point, and generates a simulation image of the robot model captured by a virtual camera that is arranged so that a position and pose of the virtual camera relative to the robot model in the virtual space coincide with the estimated position and pose of the imaging device. The control device determines an amount of correction of a position and pose of the robot for the teaching point so that the position and pose of the robot on the actual image captured after the robot has been driven according to a movement command to the teaching point approximate to the position and pose of the robot model on the simulation image.