Stereo Vision Sensor Misalignment Correction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In augmented reality (AR) systems using computer stereo vision, misalignment of optical sensors due to physical manipulation or wear can lead to inaccurate depth data, causing virtual content to be presented at incorrect locations.

Innovation Solution

A misaligned vantage point mitigation system that determines the variance in feature locations between expected and actual images from two optical sensors, calculates an adjustment variable to account for misalignment, and uses this to refine depth data calculation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If optical sensors are used for stereo vision in AR systems, then depth data can be extracted and virtual content can be presented, but sensor misalignment due to physical manipulation or wear causes inaccurate depth data and incorrect virtual content placement

Engineering Contradiction:

Improveaccuracy of depth dataVSAvoidsensor misalignment

Core Design Contradiction:

ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The system performs preliminary detection of misalignment by comparing actual feature locations with expected locations before using the depth data. This preliminary check allows the system to identify and correct misalignment issues before they affect the accuracy of virtual content placement, thereby maintaining reliability while accounting for the harmful effect of sensor misalignment.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements feedback by continuously monitoring the alignment status of optical sensors through feature location comparison. When misalignment is detected, the system adjusts the depth calculations accordingly. This feedback mechanism ensures that inaccurate depth data caused by sensor misalignment is corrected, maintaining the accuracy required for proper virtual content overlay.

Inventive Principle:

Principle #23Feedback

2Manufacturing precision

If virtual content is overlaid on real-world objects, then augmented reality effect is achieved, but misalignment causes virtual content to be presented at incorrect locations

Engineering Contradiction:

Improveprecision of virtual content placementVSAvoidaccuracy of depth data

Core Design Contradiction:

Manufacturing precisionVSReliability

Solution Approach 1:

The system replaces direct mechanical alignment dependence with a computational correction approach. Instead of relying on physical precision of sensor placement, the system uses image processing to detect feature locations, compares them with expected locations, and calculates correction factors. This substitution allows high precision virtual content placement even when sensors are physically misaligned, as the computational method compensates for the physical imperfections.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system changes the parameters used in depth calculation by introducing correction factors based on detected misalignment. When features are found at different locations than expected, the system adjusts the depth parameters accordingly before rendering virtual content. This parameter adjustment ensures that virtual content is placed accurately even when the underlying sensor geometry deviates from the assumed model.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20250008068A1Misaligned vantage point mitigation for computer stereo vision

Publication Date: 2025.01.02 SNAP INC

AI summary

Disclosed are systems, methods, and non-transitory computer-readable media for misaligned vantage point mitigation for computer stereo vision. A misaligned vantage point mitigation system determines whether vantage points of the optical sensors are misaligned from an expected vantage point and, if so, determines an adjustment variable to mitigate the misalignment based on the location of matching features identified in images captured by both optical sensors.