AR Image Synthesis via Depth-Aware Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing augmented reality technologies fail to accurately determine the positional relationship between virtual and real-world objects, leading to inconsistencies that can give users a wrong impression, as they necessarily display virtual objects on the front side regardless of their actual position in the captured image.
Innovation Solution
An image generation system that acquires captured images and depth information, determines the positional relationship between photographic objects and virtual objects in the depth direction, and synthesizes the virtual object with the captured image to ensure accurate placement based on depth information, preventing incorrect positional displays.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If the virtual object is necessarily displayed on the front side by superimposing it on the captured image, then the virtual object can be easily synthesized with the captured image, but the positional relationship between the virtual object and other objects in the captured image becomes inconsistent, giving users a wrong impression
Solution Approach 1:
The patent changes the parameter of virtual object display position from a fixed front-side position to a dynamically adjustable position based on depth information. By acquiring depth information about photographic objects and comparing it with the virtual object's depth, the system determines the appropriate display position (front or rear) to maintain consistent spatial relationships, thus resolving the contradiction between easy synthesis and positional consistency
Solution Approach 2:
The patent introduces depth information as an additional dimension for determining virtual object placement. Instead of relying solely on 2D image superposition, the system uses 3D depth data to establish accurate spatial relationships between virtual and real objects, enabling the virtual object to be correctly positioned in front of or behind photographic objects based on their relative depths
2Measurement precision
If depth information acquisition and processing are added to determine positional relationships, then the positional accuracy between virtual and real objects is improved, but the system complexity increases
Solution Approach 1:
The patent uses depth information as an intermediary element to bridge the virtual and real worlds. By acquiring depth data about photographic objects and using it as a reference for positioning virtual objects, the system achieves accurate spatial relationships without requiring complex real-time 3D reconstruction or sophisticated algorithms, thus improving positional accuracy while controlling system complexity
Data Source
AI summary
An image generation system includes a captured image acquisition section that acquires a captured image captured by an imaging section, a depth information acquisition section that acquires depth information about a photographic object observed within the captured image, an object processing section that performs a process that determines a positional relationship between the photographic object and a virtual object in a depth direction based on the acquired depth information, and synthesizes the virtual object with the captured image, and an image generation section that generates an image in which the virtual object is synthesized with the captured image.