3D Sensor Registration for Autonomous Manipulators Using Depth Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The registration of volumetric sensors with autonomous manipulators is challenging due to the need for accurate transformation between 3D sensor Cartesian Coordinates and incompletely specified 2D sensor Projective Coordinates, which is computationally burdensome and prone to measurement noise, especially in uncontrolled environments, and existing methods often require 2D image data or fragile printed patterns that decay quickly.
Innovation Solution
A method and apparatus for registering a 3D sensor with an autonomous manipulator using simple 3D shape primitives as registration targets, relying solely on depth images to obtain extrinsic registration parameters without 2D image data, ensuring independence from pre-integrated 2D sensors and maximizing signal-to-noise ratio, with a robust algorithm that delivers asymptotically optimal uncertainty.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If 2D image data and printed patterns are used for registration, then registration can be performed, but the printed patterns decay quickly and require frequent replacement
Solution Approach 1:
The patent replaces fragile printed patterns with durable 3D geometric primitives that can withstand repeated use. The 3D targets are made of robust materials and have simple geometric shapes that are easy to manufacture and replace, thereby reducing maintenance costs and extending operational lifespan.
Solution Approach 2:
The patent transitions from 2D printed patterns to 3D geometric primitives for registration. This dimensional change allows the use of depth images instead of relying on 2D image data, providing more robust and maintainable registration targets that are insensitive to lighting conditions and pattern degradation.
2Measurement precision
If 2D image data is used for registration, then registration can be performed, but it increases computational burden and is prone to measurement noise
Solution Approach 1:
The patent extracts and utilizes only the depth information from the sensor data, discarding the need for 2D image data. This extraction of the essential 3D geometric information simplifies the registration process by focusing on depth maps alone, thereby reducing computational complexity while maintaining or improving registration precision.
3Measurement precision
If complex registration procedures are used to achieve high accuracy, then registration precision improves, but the process becomes too slow for many assembly and materials handling tasks
Solution Approach 1:
The patent employs a simplified registration procedure that uses only depth images and 3D geometric primitives, performing partial registration that is sufficient for assembly and materials handling tasks. This partial action approach achieves adequate precision without the computational overhead of more complex procedures, thereby improving registration speed and productivity.
Data Source
AI summary
A method and system for registering a 3D sensor with an autonomous manipulator is provided. The 3D sensor has a field of view and a sensor coordinate system. The autonomous manipulator is a vision-guided manipulator having a work envelope and a manipulator coordinate system. The method includes moving a registration target relative to the sensor in the field of view of the sensor in the work envelope to obtain a plurality of depth maps or images of the target. The depth maps or images are processed to obtain a plurality of extrinsic registration parameters between the manipulator and the sensor.