3D Model Texturing via 2D Image Projection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for texturing 3D digital models often result in visual artifacts due to not accounting for the actual color of objects between point cloud points, requiring increased point density and powerful graphics cards for accurate representation.
Innovation Solution
A method involving meshing a point cloud into surface elements, unfolding the 3D model into a 2D plane mesh, identifying and selecting images from a bank with a normal direction close to the surface elements, and associating texture properties with corresponding polygons to produce a textured 2D model, which is then matched to create a detailed and faithful 3D model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of cloud points is increased to reduce visual artifacts, then the accuracy of color representation is improved, but the number of surface elements increases substantially exponentially, requiring powerful and expensive graphics cards
Solution Approach 1:
The patent uses 2D images from an image bank as copies of the 3D surface elements. Instead of increasing point density, the method projects 2D image textures onto the 3D mesh, allowing accurate color representation without exponentially increasing the number of surface elements. This copying approach maintains visual fidelity while avoiding the computational burden of high-density point clouds.
2Reliability
If the number of cloud points is increased to reduce visual artifacts, then the visual fidelity of the digital model is improved, but the computational resources and processing power required increase substantially
Solution Approach 1:
The method copies 2D image data onto 3D surface elements through texture mapping. This approach achieves high visual fidelity by using pre-captured image data rather than increasing point cloud density, thereby avoiding the exponential increase in computational resources required for processing and displaying high-density 3D models.
3Productivity
If linear interpolation is used to assign colors to surface elements, then the texturing process is simple and fast, but visual artifacts are produced due to not accounting for actual object colors between points
Solution Approach 1:
Instead of using linear interpolation that creates visual artifacts, the patent copies actual 2D image data onto the 3D surface elements. This preserves both the simplicity and speed of the process while eliminating visual artifacts by using real captured image data rather than mathematical approximations.
Solution Approach 2:
The patent replaces the mechanical interpolation process with an image projection process. Rather than calculating intermediate colors mathematically, the method projects actual 2D image textures onto the 3D mesh, substituting a computationally efficient image mapping operation for the artifact-prone interpolation process.
Data Source
AI summary
Method for texturing a 3D model of at least one scene (5), comprising:a) the meshing with surface elements (50; 55) of a point cloud (45) representing the scene, so as to generate the 3D model, each surface element representing an area of the scene,b) the unfolding of the 3D model for obtaining a 2D model formed of a plane mesh (60a; 60b) formed of polygons (65), each surface element corresponding to a single polygon, and vice versa, andc) for at least one, preferably all the surface elements,iv) the identification, from an image bank (40a; 40b), of the images representing the area of the scene and which have been acquired by a camera the image plane (72a-b) of which has a normal direction, in the corresponding acquisition position, forming an angle (θa-b) less than 10°, preferably less than 5°, better less than 3° with a direction normal (70) to the face of the surface element,v) the selection of an image (40a-b) from the identified images, and,vi) the association of a texture property with a corresponding polygon (65), from a piece of information of a pixel (80; 85) of the selected image which is superimposed on the surface element (55), so as to produce a textured 2D model, andd) the production of the textured 3D model by matching the 3D model and the textured 2D model.