Virtual Object Construction Using Plane Clustering for XR

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for constructing virtual objects in Extended Reality (XR) environments lack efficiency in obtaining and processing depth information from environmental images to accurately identify corner points on target objects, leading to increased computational load and reduced reconstruction accuracy.

Innovation Solution

A method that involves obtaining multiple planes from environmental images using depth information, clustering these planes to reduce data processing, and identifying corner points by analyzing the intersection lines and angles to construct a virtual object corresponding to the target object, thereby reducing data processing pressure and enhancing accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple planes are extracted from environmental images without clustering, then more corner points can be identified, but data processing pressure increases

Engineering Contradiction:

Improvecorner point identification accuracyVSAvoiddata processing pressure

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the environmental image into multiple planes first, then clusters similar planes together to reduce the number of objects needing corner point detection. This segmentation approach maintains comprehensive corner point identification while reducing overall processing complexity.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent merges similar planes through clustering based on normal vectors and other geometric features. By combining redundant plane information, the system reduces the number of planes requiring corner point extraction, thereby reducing data processing pressure while maintaining detection accuracy.

Inventive Principle:

Principle #5Merging (Combining)

2Manufacturing precision

If traditional methods are used to obtain depth information and construct virtual objects, then reconstruction can be performed, but computational load increases and accuracy decreases

Engineering Contradiction:

Improvevirtual object reconstruction accuracyVSAvoidcomputational load

Core Design Contradiction:

Manufacturing precisionVSPower

Solution Approach 1:

The patent performs preliminary clustering of planes before corner point identification and virtual object construction. This preliminary organization of data structures reduces the computational complexity of subsequent operations, lowering overall computational load while improving reconstruction accuracy through better-organized geometric data.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS11741678B2Virtual object construction method, apparatus and storage medium

Publication Date: 2023.08.29 QINGDAO PICO TECH CO LTD

AI summary

The present disclosure relates to a virtual object construction method and apparatus, and the method includes: obtaining a plurality of first planes on a target object in an environmental image according to an acquired environmental image; obtaining a plurality of corner points according to the plurality of first planes; constructing a virtual object corresponding to the target object according to a corner point located on the target object in the plurality of corner points.