Multi-Camera VR Display Control with Viewing-Driven Overlap Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods struggle to create seamless VR images using multiple cameras due to misalignment issues, especially when capturing objects at close distances, leading to visual discomfort and requiring time-consuming manual stitching processes.
Innovation Solution
A display control system that dynamically switches between live-action images captured by multiple cameras with different imaging directions, arranging them to partially overlap and adjust based on the user's viewing direction, eliminating the need for stitching and reducing misalignment-related discomfort.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of moving object
If multiple cameras are arranged to capture images in divided areas, then the viewing angle and stereoscopic visualization are improved, but misalignment occurs in boundary regions resulting in visible seams
Solution Approach 1:
The imaging area is divided into multiple regions captured by different cameras, with each camera responsible for a specific angular range. This segmentation allows wide-angle coverage while managing the complexity of image integration through defined overlapping regions between adjacent cameras.
Solution Approach 2:
An overlap area is introduced as an intermediary region between images from adjacent cameras. This overlap area contains corresponding pixel regions from both cameras, serving as a buffer zone where stitching operations can be performed to align images and eliminate visible seams at boundaries.
2Manufacturing precision
If manual stitching process is performed to create seamless images, then image alignment precision is improved, but the time and complexity of the process increases
Solution Approach 1:
The imaging system is pre-configured with cameras positioned at specific angles and overlapping fields of view, establishing the geometric relationships between cameras before image capture. This preliminary arrangement facilitates automated stitching by pre-defining the overlap areas and corresponding pixel regions that need to be aligned.
Solution Approach 2:
The manual stitching process is replaced with automated image processing algorithms that operate on the overlap areas between adjacent camera images. The system automatically identifies corresponding pixel regions in the overlap areas and performs alignment transformations, eliminating the need for time-consuming manual intervention while maintaining high precision.
3Length of moving object
If cameras are positioned close to capture short-distance objects, then the ability to image close-range targets is improved, but the degree of misalignment between adjacent images increases
Solution Approach 1:
The system dynamically adjusts imaging parameters based on object distance. For close-range objects, cameras are positioned with larger angular separations to capture the object from multiple perspectives, while the overlap area stitching process compensates for the increased misalignment through automated geometric transformation and pixel registration algorithms.
Data Source
AI summary
A viewing direction specifying unit specifies a viewing direction as a direction of a user's line of sight for the VR space. An image generation unit generates the VR image according to the viewing direction by arranging in the VR space, a plurality of live-action images captured by a plurality of cameras having different imaging directions of imaging a real space so as to have an overlap area in which the respective viewing regions of the adjacent live-action images partially overlap with each other. Further, the image generation unit dynamically switches based on the viewing direction, the live-action image to be displayed in the overlap area between the adjacent live-action images.