Depth Information Fusion for Occlusion Handling in AR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for synthesizing virtual and real objects in augmented reality require significant computational resources, making them unsuitable for terminal devices like mobile phones and tablets, which struggle with efficient occlusion handling.
Innovation Solution
A method and apparatus that utilize depth information fusion to accurately superimpose virtual objects onto real-time images captured by terminal devices, determining occlusion relationships based on depth information of both the target and virtual objects to reduce computational load and enhance AR applications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to handle occlusion relationships between virtual and real objects, then occlusion accuracy is improved, but computational load increases significantly
Solution Approach 1:
The patent segments the scene into multiple depth layers based on depth information from the depth map. By dividing the complex occlusion problem into layered segments, the system can efficiently determine occlusion relationships without performing exhaustive calculations across all objects, thus maintaining accuracy while reducing computational load.
Solution Approach 2:
The patent performs preliminary depth information acquisition and layer segmentation before the actual occlusion rendering. By pre-processing the depth data and organizing objects into depth-based layers in advance, the system avoids performing complex occlusion calculations during real-time rendering, significantly reducing computational requirements while maintaining occlusion accuracy.
2Manufacturing precision
If complex occlusion handling algorithms are implemented, then rendering accuracy is improved, but processing speed decreases
Solution Approach 1:
The patent divides the rendering process into depth-based layers, processing only the necessary layers for each frame. This segmentation allows the system to maintain high rendering accuracy through proper occlusion handling while improving processing speed by avoiding unnecessary calculations on already-resolved depth relationships.
Solution Approach 2:
The patent changes the parameter representation from complex occlusion matrices to simplified depth layer indices. By representing occlusion relationships through depth layer parameters rather than full occlusion calculations, the system achieves both accurate rendering and improved processing speed through efficient parameter comparison.
3Measurement precision
If depth information from multiple sources is integrated, then depth accuracy is improved, but device complexity increases
Solution Approach 1:
The patent merges depth information from multiple sources (depth map, image data, and virtual object depth) into a unified depth layer structure. By combining these sources into a single integrated depth representation, the system achieves high depth accuracy while managing device complexity through a unified processing framework rather than separate handling of each depth source.
Data Source
AI summary
A method for synthesizing virtual and real objects, includes: acquiring a target object in an image captured by a terminal device; in a depth information fusion mode, acquiring the depth information of the target object by a camera apparatus of the terminal device, wherein the depth information fusion mode is configured to superimpose a virtual object generated by computer modeling on the image, and the depth information of the target object is configured to indicate a distance between the target object and the terminal device; acquiring the depth information of the virtual object; and superimposing the virtual object on the image based on the depth information of the target object and the depth information of the virtual object.