3D Model Reconstruction Using Laser Pose Estimation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current computer vision methods for 3D model reconstruction in VR/AR systems face challenges in accurately capturing and reconstructing objects in real-world environments, particularly when the device moves around the target object, leading to inaccuracies and inefficiencies in pose estimation and model generation.

Innovation Solution

The method involves using laser emitters to provide laser beams, a depth camera to capture depth data, and light sensors to detect these beams for obtaining an initial camera pose, which is then used by a processing circuit to perform 3D reconstruction of the target object, employing lighthouse tracking and Iterative Closest Point (ICP) algorithms for accurate pose estimation and model building.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of stationary object

If computer vision methods are used for 3D model reconstruction when the device moves around the target object, then the coverage and completeness of the reconstructed model is improved, but the accuracy of pose estimation and reconstruction precision deteriorate

Engineering Contradiction:
Improvecoverage of reconstructed modelVSAvoidaccuracy of pose estimation
Core Design Contradiction:
Area of stationary objectVSMeasurement precision

Solution Approach 1:

The patent introduces laser beams as an intermediary reference system between the moving device and the target object. The laser emitters project structured light patterns that serve as a stable reference framework, while light sensors detect these patterns to calculate device pose. This intermediary laser reference system decouples the pose estimation accuracy from the device's movement, allowing comprehensive coverage while maintaining precision.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary pose estimation using laser beam detection before conducting the full 3D reconstruction process. By pre-establishing the device's position and orientation through laser pattern recognition, the system creates an accurate initial pose estimate that guides subsequent depth data capture, ensuring both comprehensive coverage and high reconstruction precision.

Inventive Principle:
Principle #10Preliminary action

2Device complexity

If traditional computer vision methods are used for pose estimation during device movement, then the system complexity is reduced, but the reconstruction accuracy and reliability deteriorate

Engineering Contradiction:
Improvesystem complexityVSAvoidreconstruction accuracy
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent introduces laser beams as an intermediary reference system between the moving device and the target object. The laser emitters project structured light patterns that serve as a stable reference framework, while light sensors detect these patterns to calculate device pose. This intermediary laser reference system decouples the pose estimation accuracy from the device's movement, allowing comprehensive coverage while maintaining precision.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system changes the reference parameter from visual features in the scene to the controlled laser beam pattern. By using the known, stable geometry of projected laser lines as the reference instead of relying on natural image features, the system achieves higher reliability in pose estimation while adding only moderate complexity through the laser projection and detection components.

Inventive Principle:
Principle #35Parameter changes

3Area of stationary object

If the device moves around the target object to capture comprehensive depth data, then the completeness of the 3D model is improved, but the time consumption increases

Engineering Contradiction:
Improvecompleteness of 3D modelVSAvoidtime consumption for reconstruction
Core Design Contradiction:
Area of stationary objectVSLoss of time

Solution Approach 1:

The system performs preliminary pose estimation using laser beam detection before conducting the full 3D reconstruction process. By pre-establishing the device's position and orientation through laser pattern recognition, the system creates an accurate initial pose estimate that guides subsequent depth data capture, ensuring both comprehensive coverage and high reconstruction precision.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system uses feedback from light sensor detection of laser patterns to continuously update and refine pose estimation during device movement. This real-time feedback mechanism allows the system to adapt to movement and maintain accurate reconstruction, reducing the need for multiple passes and minimizing total time consumption while ensuring model completeness.

Inventive Principle:
Principle #23Feedback

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach enhances the accuracy and speed of 3D model reconstruction, reducing time consumption and improving user experience in VR/AR environments by providing a precise and efficient method for object recognition and integration in mixed reality systems.

Implementation Method 1

capturing, by a depth camera on an electronic device, a depth data of a target object

Methodology Applied
Scientific EffectTime of flight: Time of Flight

Implementation Method 2

detecting, by one or more light sensors on the electronic device, the laser beams emitted by the one or more laser emitters

Methodology Applied
Scientific EffectLight detection: Photoelectric Effect

Data Source

PatentEP3392834B13D model reconstruction method, electronic device, and non-transitory computer readable storage medium
Publication Date: 2019.12.25 HTC CORP
  • EP3392834B1 patent drawingFigure 1
  • EP3392834B1 patent drawingFigure 2
  • EP3392834B1 patent drawingFigure 3A

AI summary

A 3D model reconstruction method includes: providing, by one or more laser emitters, laser beams; capturing, by a depth camera on an electronic device, a depth data of a target object when that the electronic device moves around the target object; detecting, by one or more light sensors on the electronic device, the laser beams emitted by the one or more laser emitters to obtain a camera pose initial value of the depth camera accordingly; and performing, by a processing circuit, a 3D reconstruction of the target object using the depth data based on the camera pose initial value to output a 3D model of the target object.