Robot Self-Position Estimation Using Panorama Structure Images

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing self-position estimation techniques for mobile robots require pre-created landmark maps, constant landmark tracking, and specialized imaging equipment, which can fail if landmarks are missed and increase operational complexity.

Innovation Solution

A self-position estimation method using captured structure images and panorama compositing, without requiring landmarks or specialized imaging, by calculating relative and absolute position coordinates based on distance information and transformation coefficients from camera parameters.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If landmark maps are created in advance and landmarks are constantly tracked for self position estimation, then self position estimation can be performed, but the system complexity increases and estimation fails when landmarks are missed

Engineering Contradiction:
Improveself position estimation reliabilityVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The invention extracts and removes the landmark dependency from the self-position estimation system. Instead of requiring pre-created landmark maps and constant landmark tracking, the system uses only the camera images captured during the robot's normal movement and work operations. This extraction of the landmark requirement simplifies the system while maintaining estimation reliability through alternative image-based positioning methods.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The camera serving the dual purpose of both work-related imaging (inspecting structures) and self-position estimation is a universal application principle. The same camera used for the robot's primary inspection tasks is also utilized for position estimation, eliminating the need for specialized imaging equipment and reducing system complexity while maintaining reliability.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Measurement precision

If pre-created landmark maps and specialized imaging equipment are used for self position estimation, then position estimation can be performed, but imaging and equipment needs increase beyond work purposes

Engineering Contradiction:
Improveself position estimation precisionVSAvoidimaging equipment complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The invention applies universality by making the camera serve dual purposes: it is used both for the robot's work-related inspection tasks and for self-position estimation. This eliminates the need for specialized imaging equipment beyond what is already required for the robot's primary function, thereby maintaining measurement precision without increasing device complexity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system uses the camera images that the robot captures during its normal work operations to simultaneously perform both the inspection work and the self-position estimation. The robot's own operational images serve the dual function, eliminating the need for separate imaging equipment or pre-created maps, thus maintaining precision while reducing equipment complexity.

Inventive Principle:
Principle #25Self-service

3Reliability

If constant landmark tracking is performed for self position estimation, then position can be estimated, but imaging operations beyond work purposes are required

Engineering Contradiction:
Improveself position estimation reliabilityVSAvoidwork efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The camera is used universally for both work-related imaging and self-position estimation simultaneously. The same imaging operations performed for inspection purposes also provide data for position estimation, eliminating the need for separate constant landmark tracking operations and thereby maintaining reliability while improving productivity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system maintains continuous self-position estimation by utilizing the continuous stream of images captured during normal work operations. Instead of requiring separate continuous landmark tracking, the useful imaging action for work purposes continuously provides data for position estimation, maintaining reliability without reducing productivity.

Inventive Principle:
Principle #20Continuity of useful action

Data Source

PatentEP3525169B1Self position estimation device, self position estimation method, program, and image processing device
Publication Date: 2026.02.18 FUJIFILM CORP
  • EP3525169B1 patent drawingFigure 1
  • EP3525169B1 patent drawingFigure 2
  • EP3525169B1 patent drawingFigure 3

AI summary

Disclosed are a self position estimation device, a self position estimation method, a program, and an image processing device that suppress imaging and movement other than a work purpose of a robot, do not need a landmark, and do not need special imaging equipment for application of the robot other than the work purpose. A self position estimation device (400) includes a first structure image acquisition unit (401) that acquires a plurality of structure images including a first structure image and a second structure image, a panorama composition unit (403) that generates a panorama composite image by subjecting the plurality of structure images including the first structure image and the second structure image to panorama composition, a first image coordinate acquisition unit (405) that acquires second image coordinates as coordinates of a specific position of the second structure image, and a first relative position coordinate calculation unit (407) that calculates relative position coordinates as relative actual coordinates of a second position as a relative self position using a transformation coefficient for transformation from an image coordinate system to an actual coordinate system.