AR Object Tracking via 3D Model Mapping and Feature Points
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current augmented reality systems face challenges in accurately superimposing computer-generated content onto real-world objects, particularly in tracking relative motion and orientation, and in providing interactive and informative AR content that is contextually relevant.
Innovation Solution
The system generates a 3D graphical model of an object, which is mapped onto its image, allowing for real-time interaction and information retrieval based on user selections, using techniques such as ray tracing and sensor data integration to track movement and orientation, and enabling the display of relevant AR content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If computer-generated content is superimposed onto real-world objects in augmented reality, then informative AR content is provided, but accurate tracking of relative motion and orientation becomes challenging
Solution Approach 1:
The patent uses feature points as intermediary elements between the real-world object and the virtual content. These feature points are detected in the image and used to establish correspondence with the 3D model, enabling accurate tracking of relative motion and orientation without directly tracking the entire object or content
Solution Approach 2:
The patent replaces traditional mechanical or direct visual tracking methods with a computational approach using feature point detection, image processing, and 3D model matching. This substitution enables more precise tracking of relative motion and orientation through algorithmic computation rather than direct mechanical measurement
2Manufacturing precision
If a 3D graphical model is mapped to an object in an image, then accurate AR overlay is achieved, but device complexity increases
Solution Approach 1:
The patent segments the complex task of AR overlay into distinct steps: detecting feature points in the image, matching these points to the 3D model, calculating relative pose, and finally rendering the overlay. This segmentation makes the complex process more manageable and implementable
Solution Approach 2:
The patent creates a digital copy (3D graphical model) of the real-world object and maps it to the detected object in the image. This copying approach enables accurate AR overlay by comparing the known 3D model with the 2D image features, simplifying the mapping process
3Ease of operation
If interactive AR content is provided based on user selections, then user interaction is enhanced, but processing time and system response delay increase
Solution Approach 1:
The patent performs preliminary actions by pre-detecting feature points and pre-establishing the 3D model before user interaction occurs. When a user selects content, the system can quickly retrieve and display information without needing to perform complex image processing at that moment, reducing response time
Data Source
AI summary
An example method is performed by a computing system, and includes: obtaining an image of an object captured by a device during relative motion between the object and the device; determining a location of the device relative to the object during image capture based on one or more attributes of the object in the image; mapping a three-dimensional (3D) graphical model representing the object to the object in the image based, at least in part, on the location of the device, where the 3D graphical model includes information about the object; receiving a selection of a part of the object; and outputting, for rendering on a user interface, at least some information from the 3D graphical model based on the part selected.