Target-Free RGBD Camera Calibration for Robotic Arms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing robotic vision systems face challenges in accurately calibrating camera extrinsic parameters relative to a robot without a separate calibration target and require precise orientation of the target, which is difficult and time-consuming.
Innovation Solution
A robotic system that uses an RGBD camera and a controller to calibrate the camera by instructing the robot to engage in various poses, extracting robot poses in the camera reference frame, and creating a transformation to align camera points with robot points without a calibration target, allowing for rapid and efficient estimation of camera parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a planar target with known appearance is placed in view of the camera for calibration, then camera intrinsic and extrinsic parameters can be inferred, but the setup process becomes time-consuming and difficult due to precise orientation requirements
Solution Approach 1:
The patent removes the calibration target from the system entirely. Instead of using a separate planar target with known appearance, the robot itself becomes the calibration object. The robot arm moves to known poses within its reference space, and the camera captures images of the robot at these poses. This extraction of the target eliminates the time-consuming setup and precise orientation requirements while maintaining calibration accuracy.
Solution Approach 2:
The robot serves multiple functions: it is both the object being observed by the camera and the calibration standard. By using the robot's own known poses and reference frame as the calibration basis, the system eliminates the need for a separate calibration target. The robot arm's ability to precisely reproduce known poses provides the necessary geometric reference for calibration without requiring external equipment.
2Measurement precision
If a calibration target is used to determine camera extrinsic parameters, then accurate camera-to-robot transformation can be achieved, but the device complexity increases due to additional calibration equipment and procedures
Solution Approach 1:
The calibration target is completely removed from the system. The patent uses only the robot arm and camera that are already part of the robotic system. By extracting the separate calibration target and using the robot itself as the calibration reference, the system reduces device complexity while maintaining transformation accuracy through the robot's known reference frame and precise pose reproduction capability.
3Measurement precision
If traditional camera calibration methods are used with a checkboard pattern and corner detector, then intrinsic parameters can be determined, but the method cannot easily provide camera-to-robot transformation when the target orientation relative to the robot is unknown
Solution Approach 1:
The robot serves as both the observed object and the reference frame for calibration. By capturing images of the robot at known poses within its reference space, the system simultaneously determines camera intrinsic parameters and establishes the camera-to-robot transformation. The robot's reference frame, defined relative to its base, provides the necessary geometric reference that enables both functions without requiring separate calibration targets or known target orientations.
Data Source
AI summary
One embodiment provides a robotic system comprising: a robot, the robot further comprising a moveable robotic arm that moves within the robot's reference space; a depth-sensing camera, the camera having a reference frame that is in substantial view of the robot's reference space; a controller, the controller further comprising a processor and a computer readable memory that comprises instructions such that, when read by the controller, the controller inputs image data from the camera and sends signals to the moveable robotic arm, the instructions further comprising the steps of: calibrating the camera to the robot by instructing the robot to engage in a number of robot poses; extracting the location of the robot poses to obtain the robot poses in the camera reference frame; and creating a transformation that transforms robot points afterwards to camera points.