Automated 3D Texture Mapping via Perspective Splicing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D texture mapping methods are labor-intensive and time-consuming, with low accuracy due to the difficulty in manually aligning and splicing texture images from different perspectives, especially in augmented and virtual reality applications.
Innovation Solution
An automated image processing method and apparatus that acquire a 3D model and original texture images, determine the mapping relationship, select a subset of texture images based on a specific perspective, splice them into a coherent image, and map it back to the 3D model, using techniques like neural networks and image segmentation to optimize sharpness and integrity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual methods are used for 3D texture mapping, then flexibility and control are maintained, but accuracy and time efficiency deteriorate
Solution Approach 1:
The patent replaces manual mechanical operations with automated image processing systems. Specifically, it uses perspective transformation algorithms, image splicing techniques, and neural network-based optimization to automatically perform texture mapping tasks that were previously done manually, thereby improving both accuracy and efficiency
Solution Approach 2:
The system enables self-service automation where the computer automatically performs texture mapping operations without human intervention. The automated workflow includes acquiring images, determining perspectives, splicing images, and mapping textures to 3D models autonomously, eliminating the need for manual alignment and splicing while maintaining high precision
2Productivity
If manual alignment and splicing of texture images is performed, then control over details is maintained, but productivity and accuracy deteriorate
Solution Approach 1:
The patent substitutes manual alignment and splicing operations with automated computer-based image processing. It employs perspective transformation mathematics, feature point detection algorithms, and neural networks to automatically align and splice texture images, achieving both high productivity and precise alignment that cannot be consistently achieved manually
Solution Approach 2:
The system creates accurate digital copies and representations of the object from multiple perspectives, then uses these copies to generate the final spliced texture image. This copying approach allows for precise digital manipulation and alignment without the limitations of manual physical alignment
3Reliability
If multiple texture images from different perspectives are spliced, then completeness of the 3D model is improved, but complexity of the process increases
Solution Approach 1:
The patent segments the texture mapping process into distinct automated stages: image acquisition, perspective determination, image splicing, and final mapping. By dividing the complex task of combining multiple perspective images into manageable automated segments, the system achieves complete 3D texture coverage while keeping each processing step manageable and controllable
Solution Approach 2:
The system replaces the complexity of manually coordinating multiple perspective images with automated computational methods. Computer algorithms automatically determine perspectives, select appropriate images, perform splicing operations, and map textures to the 3D model, transforming a complex manual coordination task into a streamlined automated process
Data Source
AI summary
The present disclosure discloses an image processing method, apparatus, and a non-transitory computer readable medium. The method can includes: acquiring a three-dimensional (3D) model and original texture images of an object, wherein the original texture images are acquired by an imaging device; determining a mapping relationship between the 3D model and the original texture images of the object; determining, among the original texture images, a subset of texture images associated with a first perspective of the imaging device; splicing the subset of texture images into a spliced texture image corresponding to the first perspective; and mapping the spliced texture image to the 3D model according to the mapping relationship.