3D Pose Estimation Using Automatic Feature Line Pairing

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for estimating the position and orientation of an imaging apparatus in three-dimensional space using augmented reality and CAD data are cumbersome, requiring manual selection of feature lines and line segments, which can be time-consuming and prone to incorrect selections, especially for inexperienced users.

Innovation Solution

An estimation apparatus that detects feature lines and points in an image, associates them with CAD data line segments, and generates corresponding pairs to estimate the imaging apparatus's position and orientation, simplifying the selection process by displaying feature lines and points for easy selection and automatically generating additional pairs to reduce estimation errors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual selection of feature lines and line segments is used for position and orientation estimation, then the estimation can be performed using existing methods, but the process becomes time-consuming and prone to incorrect selections

Engineering Contradiction:

Improveposition and orientation estimation accuracyVSAvoidtime for selecting feature lines and line segments

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary detection of multiple candidate feature lines in the image and multiple candidate line segments in the CAD data before the actual estimation process. This preliminary preparation allows the subsequent pairing process to be more efficient and accurate, reducing the time required for manual selection while maintaining estimation precision.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system generates virtual line segments by connecting detected feature points in the image, creating additional corresponding pairs without requiring manual selection. These virtual line segments serve as copies or derivatives of the detected features, expanding the available data for estimation while reducing manual intervention time.

Inventive Principle:

Principle #26Copying

2Ease of operation

If manual selection of feature lines and line segments is used, then the existing estimation methods can be applied, but the process becomes cumbersome and error-prone for inexperienced users

Engineering Contradiction:

Improveease of selecting feature lines and line segmentsVSAvoidcorrectness of selected combinations

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

The system automatically performs the pairing between feature lines and line segments by detecting corresponding features in both the image and CAD data. This self-service approach eliminates the need for manual pairing, making the system easy to operate while ensuring reliable and correct combinations through automated feature matching algorithms.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system uses the detected feature lines and line segments to perform position and orientation estimation, then utilizes this estimation result to validate and refine the feature pairings. This feedback mechanism ensures that the selected combinations are correct and improves the reliability of the estimation process.

Inventive Principle:

Principle #23Feedback

3Measurement precision

If more feature lines and line segments are selected to improve estimation accuracy, then the precision of position and orientation estimation increases, but the complexity of the selection process increases

Engineering Contradiction:

Improveposition and orientation estimation accuracyVSAvoidcomplexity of feature selection process

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system segments the feature selection process into automatic detection of feature lines in the image and automatic extraction of line segments from CAD data. This segmentation eliminates the need for manual selection complexity while allowing multiple features to be processed systematically, improving estimation accuracy without increasing operational complexity.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system replaces the mechanical manual selection process with automated computer vision algorithms that detect feature lines and match them with CAD data line segments. This substitution eliminates the complexity of manual feature selection while enabling the processing of multiple features to improve estimation precision.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS10930068B2Estimation apparatus, estimation method, and non-transitory computer-readable storage medium for storing estimation program

Publication Date: 2021.02.23 FUJITSU LTD

AI summary

An estimation apparatus is configured to obtain shape information containing information about multiple line segments that depict a shape of an object; detect multiple feature lines in an image of the object captured by an imaging apparatus; receive a first instruction for associating a feature line selected from the multiple feature lines with a line segment selected from the multiple line segments and a second instruction for associating two points selected in the image with two end points selected from end points of the multiple line segments; generate a first line segment connecting the two points and a second line segment connecting the two end points; and estimate a position and orientation of the imaging apparatus in three-dimensional space by using a combination of the selected feature line and the selected line segment and a combination of the first line segment and the second line segment.