Virtual Object Placement From Spherical Space Structure Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for combining virtual objects with captured images often result in unnatural positioning due to the lack of accurate automatic arrangement, requiring manual input of floor plans and user intervention.
Innovation Solution
An image processing method that estimates the structure of a space from a spherical image, identifies regions for virtual object placement, and automatically arranges the object based on this structure, eliminating the need for manual input.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If virtual objects are combined with captured images using existing methods, then the combination can be achieved, but the virtual objects are arranged at unnatural positions reducing accuracy of automatic arrangement
Solution Approach 1:
The system performs preliminary structure estimation of the space from the background image before combining virtual objects. By pre-analyzing the spatial structure, the system determines appropriate arrangement regions in advance, ensuring virtual objects are placed at natural positions that conform to the actual space geometry, thereby improving both accuracy and naturalness of automatic arrangement
Solution Approach 2:
The patent introduces an intermediary structure estimation process that acts as a mediator between the background image and virtual object placement. This intermediary analysis of spatial structure provides the basis for determining natural arrangement regions, enabling accurate positioning without requiring manual intervention while ensuring objects are placed at realistic locations
Data Source
AI summary
An image processing method to be executed by an image processing system includes: estimating a structure of a space inside a construction from a background image in which the space is imaged in all directions; estimating a region in which a virtual object is allowed to be arranged in the space based on the estimated structure; and combining the virtual object with the background image in the estimated region.