2D Whole-Image Reconstruction with Pixel-Based View Alignment

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for generating composite images from individual images captured with translational camera movement suffer from reduced image quality and distortions, particularly when the camera position changes relative to the object or recording area.

Innovation Solution

A method for generating a two-dimensional composite image by determining the spatial orientation of individual images based on pixel values and optimizing the relative alignment without additional aids, using a movable recording unit to capture sub-areas from different viewpoints and distances, and projecting these images onto a common image plane to form an overall image.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of stationary object

If individual images are captured with translational camera movement to cover a large recording area, then the coverage area is improved, but image quality and resolution deteriorate due to position changes

Engineering Contradiction:
Improverecording area coverageVSAvoidimage quality and resolution
Core Design Contradiction:
Area of stationary objectVSMeasurement precision

Solution Approach 1:

The patent replaces mechanical alignment methods (physical markers, manual positioning) with a computational approach based on image data analysis. The system automatically determines spatial orientations by comparing pixel values and optimizing quality metrics, eliminating the need for mechanical reference systems and enabling high-precision alignment despite camera movement.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent changes the approach from assuming fixed camera position to actively optimizing spatial orientation parameters. By varying the relative spatial orientations as adjustable parameters and optimizing them based on pixel value comparisons, the system adapts to translational movement while maintaining image quality.

Inventive Principle:
Principle #35Parameter changes

2Ease of manufacture

If traditional panorama stitching methods are used with spherical projection, then the processing is simplified, but distortions increase when camera position changes significantly

Engineering Contradiction:
Improveprocessing simplicityVSAvoidimage distortion
Core Design Contradiction:
Ease of manufactureVSManufacturing precision

Solution Approach 1:

The patent replaces traditional spherical projection geometry with a computational optimization approach. Instead of forcing images onto a spherical surface, the system uses pixel value comparisons and quality metric optimization to determine appropriate spatial orientations, reducing geometric distortions while maintaining processing feasibility.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent introduces dynamic optimization of spatial orientations based on actual image content rather than static geometric assumptions. The relative spatial orientations are adjusted dynamically to maximize image quality metrics, allowing the system to adapt to varying camera positions and reduce distortions.

Inventive Principle:
Principle #15Dynamics

3Measurement precision

If additional aids like markers or structures are used for alignment, then alignment accuracy is improved, but device complexity increases

Engineering Contradiction:
Improvealignment accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent enables the imaging system to align images using only the image data itself, without external markers or reference structures. The system extracts alignment information from the pixel values within the images, making the alignment process self-contained and eliminating additional hardware components.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent extracts alignment information directly from the pixel values of the images themselves, removing the need for separate marker systems. By utilizing the inherent information in the image data, the system achieves accurate alignment without adding external aids.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3642848B1Method for producing a two-dimensional whole image
Publication Date: 2025.10.29 SIRONA DENTAL SYSTEMS GMBH CORP LEGAL
  • EP3642848B1 patent drawingFigure 1
  • EP3642848B1 patent drawingFigure 2~3
  • EP3642848B1 patent drawingFigure 4A~4B

AI summary

The invention relates to a method for producing a two-dimensional whole image of a recording region, which is captured by means of a plurality of individual images each having its own viewing direction and its own distance, a spatial orientation of a main image plane of each individual image relative to a main image plane of the particular further individual image being determined on the basis of an overlap of the captured partial regions, and at least a plurality of individual images being combined in accordance with the spatial orientations to form the whole image. The whole image area is the surface of a torus.