Multi-Angle Image Capture for Low-Cost 3D Object Presentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Three-dimensional modeling of objects requires high-end capturing devices, high-quality photographing environments, long reconstruction periods, and high manual post-processing costs, leading to high costs and potential reconstruction failures for reflective, transparent, and less textured articles, and NeRF solutions face long training times and low PSNR with cloud noise issues.
Innovation Solution
A low-cost three-dimensional information presentation method using common image capturing devices like mobile phones, combined with SLAM, AR, and data processing technologies, to capture and transform images into multiple viewing angles, reducing costs and improving presentation quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If three-dimensional reconstruction technology is used to present three-dimensional information, then the presentation quality is improved, but the cost and complexity increase significantly
Solution Approach 1:
The patent segments the three-dimensional information capture process into multiple two-dimensional image captures from different viewing angles. Instead of using a complex three-dimensional reconstruction system, it divides the task into capturing images at multiple positions around the article, then synthesizes these 2D images to create the 3D presentation effect.
Solution Approach 2:
The patent uses multiple two-dimensional copies (images) of the article from different angles to represent and reconstruct the three-dimensional information. By capturing and processing multiple 2D image copies, the system achieves 3D presentation without requiring complex 3D scanning hardware.
2Manufacturing precision
If three-dimensional reconstruction technology is used, then the presentation quality is improved, but the time and cost for reconstruction increase
Solution Approach 1:
The patent performs preliminary actions by capturing multiple images from different viewing angles in advance. These pre-captured images are then used to generate the three-dimensional presentation, eliminating the need for time-consuming real-time reconstruction processing.
Solution Approach 2:
By using multiple pre-captured two-dimensional image copies from different angles, the system avoids time-consuming 3D reconstruction algorithms. The presentation is generated by synthesizing these pre-existing image copies, significantly reducing processing time.
3Ease of manufacture
If conventional image capturing is used for three-dimensional reconstruction, then the equipment cost is reduced, but the reconstruction fails for reflective, transparent, and less textured articles
Solution Approach 1:
The patent changes the approach from attempting direct 3D reconstruction to capturing 2D images from multiple angular dimensions. By arranging image capturing positions at different viewing angles around the article, it captures information that conventional single-angle methods miss, improving success rate for challenging materials.
Solution Approach 2:
The patent applies local quality by using different image capturing positions with specific viewing angles tailored to capture different aspects of the article. Each capturing position is optimized for its specific angle, ensuring that reflective, transparent, and textured surfaces are all captured with appropriate lighting and perspective.
4Manufacturing precision
If NeRF solutions are used for three-dimensional presentation, then the presentation quality is improved, but the training time increases and cloud noise appears
Solution Approach 1:
Instead of using complex NeRF models that require extensive training, the patent uses a simpler, more direct approach of capturing and synthesizing images from pre-defined viewing angles. This lighter method avoids long training periods while still achieving high-quality 3D presentation.
Solution Approach 2:
The patent captures images from more viewing angles than the minimum required, ensuring comprehensive coverage of the article from all perspectives. This excessive sampling approach guarantees high-quality reconstruction without requiring complex training algorithms.
Data Source
Figure 1
Figure 2a~2b
Figure 3a~3b
AI summary
The present disclosure relates to an image processing method, apparatus and system, and relates to the technical field of image processing. The image processing method comprises: according to a position of an article undergoing acquisition, determining a three-dimensional area, the whole or part of said article being located within the three-dimensional area; providing on the surface of the three-dimensional area a plurality of image acquisition positions; and at each image acquisition position, using an image acquisition apparatus to acquire a first image of said article in a viewing angle at the image acquisition position, so as to obtain a plurality of first images of said article at different viewing angles, so as to help an image display apparatus to display, according to the plurality of first images, an image corresponding to a selected viewing angle.