3D Model Resolution Uniformity via Multi-View Camera Acquisition

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for constructing 3D digital models from images acquired by cameras result in varying pixel resolutions based on distance, leading to poorly detailed virtual objects when displayed, as objects far from the camera are rendered with lower resolution.

Innovation Solution

The method involves strategically positioning cameras at multiple acquisition locations to ensure overlapping images with high-resolution portions (>50% areal density) are captured, allowing for the creation of a 3D model with uniform resolution by focusing on these high-resolution areas and discarding lower resolution parts.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If images are acquired by a stationary camera at various locations, then the acquisition process is simple, but objects far from the camera are reproduced with lower resolution

Engineering Contradiction:
Improveimage resolutionVSAvoidacquisition process
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent applies the dynamics principle by making the acquisition device mobile rather than stationary. The device moves to multiple acquisition locations around the scene, allowing the camera to maintain optimal distances from different objects. This dynamic approach ensures that each object is captured at high resolution from appropriate viewpoints, resolving the contradiction between maintaining simple acquisition processes and achieving uniform high resolution across all objects.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent transitions from a single stationary viewpoint to multiple spatial locations, adding the dimension of spatial movement to the acquisition process. By positioning the camera at different locations around the scene, the system captures objects from optimal distances, ensuring uniform high resolution throughout the 3D model without requiring complex processing at a single viewpoint.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Area of stationary object

If the camera is positioned far from objects to capture the entire scene, then the field of view is comprehensive, but the resolution of distant objects decreases

Engineering Contradiction:
Improvefield of viewVSAvoidobject resolution
Core Design Contradiction:
Area of stationary objectVSManufacturing precision

Solution Approach 1:

The patent applies segmentation by dividing the scene into multiple regions, each captured from a specific acquisition location. Instead of attempting to capture the entire scene from a single distant position, the device moves to capture different portions of the scene up close, ensuring each region is imaged at high resolution while maintaining comprehensive scene coverage through the combination of multiple images.

Inventive Principle:
Principle #1Segmentation

3Manufacturing precision

If multiple images are acquired from different locations to improve resolution, then uniform resolution is achieved, but the number of images and processing complexity increases

Engineering Contradiction:
Improveuniform resolutionVSAvoidacquisition efficiency
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The patent applies local quality by identifying and prioritizing the capture of high-resolution portions for each object in the scene. The acquisition process focuses on obtaining sufficient high-resolution coverage for each object from appropriate viewpoints, rather than uniformly oversampling the entire scene. This approach achieves uniform resolution effectiveness while optimizing acquisition efficiency.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS11297229B2Method of acquiring images at a plurality of acquisition locations of an acquisition device
Publication Date: 2022.04.05 SOLETANCHE FREYSSINET SAS
  • US11297229B2 patent drawing
  • US11297229B2 patent drawing

AI summary

A method of acquiring images includes moving, into a plurality of acquisition locations, of an acquisition device including at least one camera, and acquisition at each acquisition location of at least one image of a scene by the camera. Each acquisition location being chosen in such a manner that scenes viewed by the camera at two consecutive acquisition locations and corresponding images overlap, at least partially, and areal density of pixels assigned to at least one element of the corresponding scene, which is represented in the corresponding image by a high-resolution portion, is greater than 50% or greater than 80% of a target areal density, the areal density of pixels being defined as a ratio of the area of the element projected in a plane perpendicular to an optical axis of the camera over a quantity of pixels of the high-resolution portion.