Mobile AR Virtual Image Positioning Using Object Anchors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing augmented reality methods require visible markers like QR codes for accurate virtual object positioning, which can be unreliable at large distances or under adverse conditions, and fail when markers disappear from view.
Innovation Solution
A method and apparatus that use a mobile device to superimpose virtual images onto real scenes by analyzing camera stream images, calculating three-dimensional positions, and employing object portions as anchor points, allowing virtual object positioning even without constant marker visibility, using sensors like acceleration and rotary sensors for relative position measurement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If marker-based positioning is used for augmented reality, then virtual object positioning accuracy is improved, but the system fails when markers are not visible or at large distances
Solution Approach 1:
The patent uses object portions (natural features in the scene) as intermediary anchor points between the camera and virtual objects. When markers are not visible, the system automatically transitions to using detected object portions from the camera stream to maintain virtual object positioning, thus resolving the contradiction between marker-based precision and reliability without constant marker visibility
Solution Approach 2:
The system dynamically changes the reference parameter from marker coordinates to object portion coordinates based on visibility conditions. When markers disappear or are at large distances, the positioning system switches to using naturally detected object features, maintaining positioning reliability across varying conditions
2Stability of the object's composition
If multiple sensors are used for position measurement, then positioning stability over wide distances is improved, but device complexity increases
Solution Approach 1:
The patent makes existing mobile device sensors (acceleration sensor, rotary sensor) perform multiple functions: they are used both for general device orientation detection and specifically for calculating relative position changes of the camera with respect to anchor points. This multi-functionality approach improves positioning stability without significantly increasing device complexity
Solution Approach 2:
The system merges data from multiple sensors (camera, acceleration sensor, rotary sensor) and processing methods (image analysis, sensor fusion) into a unified positioning system. By combining these elements, the patent achieves stable positioning over wide distances while utilizing existing device components rather than adding separate dedicated systems
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
Enables accurate and stable virtual object positioning over wide distances and varying orientations, maintaining correct perspective and positioning without relying on constant marker visibility, enhancing augmented reality applications in mobile devices.
Implementation Method 1
at least one measuring device located in the mobile device, in addition to the image acquisition of the invention, for measuring the change in the relative position is being used
Implementation Method 2
using sensors like acceleration and rotary sensors for relative position measurement
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for superimposing a virtual image onto an image of a real scene comprises a step of reading image data representing an image (108) of the real scene (106) captured by a surroundings capturing device (102) of a mobile device (100), a step of determining marker data (132) from the image data, said marker data (132) representing an image (116) and a position of a marker (112) located in the real scene (106), a step of reading virtual image data (140) representing image data selected from a plurality of virtual image data (140) using marker data (132), said virtual image data (140) comprising representation instructions for representing the virtual image (124) and position instructions for positioning the virtual image (124), a step of determining object data (134) from the image data, said object data (134) representing an image (118) and a position of an object portion (114) of an object (110) located in the surroundings of the marker (112) in the real scene (106), and a step of determining a position rule (142) for positioning the virtual image (124) in relation to the image (118) of the object portion (114) using the object data (134) and the virtual image data (140).