Marker-Free AR Component Positioning for Product Assembly
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional Augmented Reality (AR) systems face challenges in efficiently positioning virtual objects within a user's field of view, particularly in complex and changing physical environments, such as during product manufacturing and assembly, due to difficulties in marker placement and limited processing capabilities.
Innovation Solution
The system generates and displays virtual representations of components based on three-dimensional surface data of partially assembled products, calculating their location, orientation, size, and shape without relying on markers, suitable for use with modest computing resources like AR headsets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If markers are placed on physical objects for virtual object positioning, then positioning accuracy is improved, but marking difficulty increases due to changing external profiles during manufacturing/assembly
Solution Approach 1:
The patent uses a digital 3D model (copy) of the physical object instead of physical markers. The system captures the physical object's geometry and creates a digital representation that is used for virtual object positioning, eliminating the need to place physical markers on the object surface.
Solution Approach 2:
The patent replaces the mechanical marker placement system with an optical scanning and computational system. Instead of physically attaching markers to objects, the system uses cameras or sensors to scan the object's geometry and computationally determines positioning based on the captured 3D data.
2Ease of manufacture
If object recognition is used for environment tracking, then marker placement is eliminated, but processing requirements increase beyond capabilities of portable AR devices
Solution Approach 1:
The system performs preliminary scanning and 3D model creation offline or in advance. The complex processing of object recognition is done before the AR session begins, or the 3D model is pre-processed and stored, so that during actual AR operation, only lightweight matching and positioning computations are needed on the portable device.
Solution Approach 2:
The patent divides the processing into separate stages: (1) initial 3D scanning and model creation, (2) feature extraction and matching, and (3) real-time positioning. This segmentation allows heavy computational tasks to be performed offline or on more powerful systems, while the portable AR device only handles the lighter real-time positioning computations.
Data Source
AI summary
Systems and methods include determination of a first component of a set of components under assembly in a physical environment, determination of a first physical position of a user with respect to the first component in the physical environment, determination of a second component of the set of components under assembly based on assembly information associated with the set of components, determination of three-dimensional surface data of the second component, determination of a physical relationship between the first component and the second component based on a model associated with the set of components, determination of a graphical representation of the second component based on the first physical position of the user with respect to the first component, the physical relationship between the first component and the second component, and the three-dimensional surface data of the second component, and presentation of the graphical representation to the user in a view including the first component in the physical environment, wherein the presented graphical representation appears to the user to be in the physical relationship to the first component.