3D Model Generation System for Complex Spaces

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing techniques for generating 3D models of spaces are limited in accurately estimating complex shapes, such as rooms with pillars, as they are primarily designed for cuboid shapes and struggle with overlapping surface captures.

Innovation Solution

A 3D model generating system that splits panorama images into overlapping partial models, determines their positions and orientations based on image information, and combines them to create an overall 3D model of the space, even with multiple orthogonal surfaces.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If techniques are used to generate a 3D model of a cuboid-shaped room based on panorama images, then the 3D model generation is simple and efficient, but the accuracy of estimating complex space shapes (such as rooms with pillars) deteriorates

Engineering Contradiction:
Improve3D model generation efficiencyVSAvoidspace shape estimation accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent divides the space into multiple regions (first region and second region) based on different capturing directions. Each region is processed independently to generate separate 3D models, which are then combined. This segmentation allows the system to handle complex shapes with pillars and irregular structures while maintaining efficient processing for each simpler sub-region.

Inventive Principle:
Principle #1Segmentation

2Area of stationary object

If a single 3D model is generated for the entire space, then the model represents the complete space, but the handling of overlapping surfaces and complex geometries becomes difficult

Engineering Contradiction:
Improveoverall space coverageVSAvoidmodel processing complexity
Core Design Contradiction:
Area of stationary objectVSDevice complexity

Solution Approach 1:

The patent segments the overall space into multiple regions that are captured from different directions. Each region generates a separate 3D model representing a portion of the total space. This segmentation reduces the complexity of handling overlapping surfaces by treating each region independently, while still achieving complete space coverage through combination of multiple regional models.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent combines multiple partial 3D models from different regions to generate a comprehensive overall 3D model of the entire space. This merging process integrates the individual regional models while managing overlapping surfaces, achieving complete space representation without requiring all complex processing to occur on a single monolithic model.

Inventive Principle:
Principle #5Merging (Combining)

3Measurement precision

If multiple images with overlapping capturing ranges are used to generate partial models, then the coverage of complex space structures is improved, but the process of integrating these partial models into an overall model becomes more complex

Engineering Contradiction:
Improvespace shape estimation accuracyVSAvoidmodel integration complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the imaging process into multiple directional captures, where each image captures a specific region from a particular direction. The overlapping capturing ranges ensure complete coverage of complex structures like pillars. The segmentation approach allows systematic integration of these partial models by processing each region independently before combination, reducing the overall integration complexity.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3467787B1Three-dimensional model generation system, three-dimensional model generation method, and program
Publication Date: 2025.01.01 RAKUTEN GROUP INC
  • EP3467787B1 patent drawingFigure 1
  • EP3467787B1 patent drawingFigure 2
  • EP3467787B1 patent drawingFigure 3

AI summary

A 3D model generating system for accurately estimating a shape of an entire space surrounded by a plurality of surfaces is provided. Image acquiring means (101) of the 3D model generating system (1) acquires a plurality of images respectively capturing different directions in a space surrounded by a plurality of surfaces. Partial model generating means (102) generates a plurality of partial models respectively corresponding to the images, each of the partial models representing a portion of the surfaces captured in the images. Overall model generating means (103) generates an overall model representing an entire surface based on the partial models generated by the partial model generating means (102).