Equirectangular Image Generation for 3D Virtual Environments
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Rendering three-dimensional (3D) virtual environments from 3D models is computationally intensive, requiring specialized hardware and software, limiting the type of devices that can view these environments from different perspectives.
Innovation Solution
Generating synthetic equirectangular images directly from 3D virtual environments by specifying a 3D virtual camera position, acquiring image strips of pixels at multiple yaw angles, and assembling them to form a 360° view, which can be output without needing high-performance graphics cards or specialized software/hardware.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 3D virtual environments are rendered from 3D models using traditional methods, then high-quality 3D visualization is achieved, but computational intensity increases and specialized hardware is required
Solution Approach 1:
The patent creates equirectangular image copies of the 3D virtual environment that can be displayed on standard devices without requiring specialized rendering hardware. Instead of rendering the full 3D environment in real-time on the display device, the system pre-generates equirectangular images from the 3D model and stores them, then displays these copied images on modestly-equipped devices, eliminating the need for high-performance graphics cards while maintaining visualization quality
Solution Approach 2:
The patent segments the 3D virtual environment into multiple equirectangular images taken from different virtual camera positions. Each equirectangular image represents a complete 360-degree view from a specific position, allowing the environment to be broken down into discrete, pre-rendered components that can be stored and displayed independently on standard devices
2Ease of operation
If traditional 3D rendering methods are used, then interactive viewing from different perspectives is enabled, but the system requires high-performance graphics cards and specialized software
Solution Approach 1:
The system creates copies of the 3D environment in equirectangular format that can be viewed interactively on standard devices. By pre-rendering multiple equirectangular images from different virtual camera positions and allowing users to navigate between these pre-computed views, the system enables interactive viewing without requiring the display device to perform real-time 3D rendering
Solution Approach 2:
The patent replaces the mechanical 3D rendering system (requiring GPU acceleration and specialized software) with a 2D image display system. Instead of using complex graphics processing to render 3D scenes in real-time, the system substitutes this with simpler image display of pre-rendered equirectangular photos, allowing standard devices to provide interactive viewing capability
3Area of stationary object
If equirectangular images are generated from multiple virtual camera positions, then complete 360-degree coverage is achieved, but the number of images and data size increases
Solution Approach 1:
The patent uses equirectangular projection, which maps the spherical 360-degree field of view onto a rectangular 2D plane. This spherical-to-cylindrical mapping allows complete 360-degree coverage to be achieved in a single image rather than requiring multiple images, reducing the total quantity of images needed while maintaining full angular coverage
Data Source
AI summary
An equirectangular image of a three-dimensional (3D) virtual environment is generated in a computer-automated fashion. In one example, a 3D virtual position of a virtual camera in a 3D virtual environment is specified. For each of a plurality of different yaw angles rotated about an axis extending through the 3D virtual position, the virtual camera is used to acquire an image strip of pixels parallel to the axis of rotation. Image strips of pixels of the 3D environment acquired at the different yaw angles are assembled to form an equirectangular image of the 3D virtual environment from the specified 3D virtual position.