Occluded Area Determination Using Feature Point Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining occluded areas of virtual objects in augmented reality require significant computational resources, especially due to the need for reconstructing three-dimensional models and real-time depth calculations.
Innovation Solution
The method constructs a three-dimensional map of a current frame using feature points and depth information, reducing the amount of image data required and improving construction speed. It also determines occluded areas in subsequent frames by matching feature points with a preset threshold, reducing the need for full-frame image data processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If three-dimensional model reconstruction method is used to determine occluded area, then occlusion accuracy is improved, but computational complexity increases significantly
Solution Approach 1:
The patent segments the image data processing by selecting only feature points from the current frame instead of processing all image data. This segmentation approach maintains occlusion accuracy by focusing on key geometric features while significantly reducing computational complexity by excluding redundant pixel-level data processing.
Solution Approach 2:
The patent extracts only the essential elements (feature points and their depth information) needed for occlusion determination from the complete image data. By taking out only the relevant feature points rather than processing the entire image, the method achieves accurate occlusion detection with reduced computational burden.
2Measurement precision
If all image data of current frame is processed to construct scene three-dimensional map, then map accuracy is improved, but processing speed decreases
Solution Approach 1:
The patent segments the image data by selecting only feature points from the current frame to construct the scene three-dimensional map. This segmentation maintains map accuracy by focusing on key geometric features while significantly improving processing speed by excluding redundant pixel-level data.
Solution Approach 2:
The patent extracts only the essential feature points and their depth information from the complete image data for map construction. By taking out only the relevant geometric features rather than processing all image data, the method achieves accurate three-dimensional mapping with improved processing efficiency.
3Measurement precision
If depth information is recalculated for each frame change, then depth accuracy is improved, but computational load increases
Solution Approach 1:
The patent performs preliminary depth calculation only when the current frame changes significantly, rather than recalculating depth information for every frame. This preliminary action approach maintains depth accuracy by calculating depth information in advance when needed, while reducing computational load by avoiding redundant recalculations during minor frame transitions.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method and apparatus for determining an occluded area of a virtual object, and a terminal device, applicable to the technical field of software. The method comprises: constructing a three-dimensional scene map of a current frame according to feature points of the current frame and corresponding depth information (S11); if a click operation of a user on the three-dimensional scene map is detected, displaying a specified virtual object on a position corresponding to the click operation (S12); constructing a three-dimensional scene model according to feature point information of the current frame (S13); comparing depth values of the three-dimensional scene model and a model of the virtual object (S14); determining an occluded area of the virtual object in the current frame according to the comparison result (S15); and determining an occluded area of the virtual object in the next frame according to the occluded area of the virtual object in the current frame (S16). The method can greatly improve the arithmetic speed of identifying the occluded area of the virtual object in the sequence frame.