Direct Plane Fitting for Fast Planar Surface Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing indirect plane fitting methods for generating 3D models from 2D images are slow, cumbersome, and inaccurate, particularly for low-texture structures and scenes with occluding objects, leading to inefficiencies and inaccuracies in modeling planar surfaces.
Innovation Solution
A direct plane fitting method that utilizes two images captured from different known positions to identify common regions, compute gradients, and project rays to intersect at a world-point, employing similarity metrics and ray tracing to determine plane vectors without intermediate feature extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If indirect plane fitting methods are used to generate 3D models from 2D images, then feature extraction and processing can be performed, but the computation is intensive and slow
Solution Approach 1:
The patent extracts only the essential geometric information needed for plane fitting directly from image pixels, bypassing the need for complex intermediate feature extraction and point cloud generation. This selective extraction of necessary data reduces computational overhead while maintaining modeling accuracy.
Solution Approach 2:
Instead of the traditional approach of extracting features from pixels and then fitting planes, the patent inverts the process by directly fitting planes to pixel coordinates through mathematical optimization, eliminating the intermediate processing steps and significantly reducing computation time.
2Reliability
If traditional indirect plane fitting methods are used, then point clouds can be generated, but the method fails for low-texture structures and occluding objects
Solution Approach 1:
The patent introduces homography matrices as an intermediary mathematical tool that maps points between different image views to 3D world coordinates. This intermediary approach allows robust plane fitting even when direct feature matching fails due to low texture or occlusions, as the homography provides a geometric constraint that guides the fitting process.
3Measurement precision
If indirect methods with multiple processing steps are used, then features can be extracted and processed, but the process is cumbersome and time-consuming
Solution Approach 1:
The patent performs preliminary computation of homography matrices and projection rays before the actual plane fitting process. These pre-computed geometric constraints are then used to guide the optimization, reducing the search space and accelerating the fitting process while maintaining precision.
Solution Approach 2:
The patent replaces the mechanical multi-step process of feature extraction, point cloud generation, and iterative fitting with a direct mathematical optimization approach that uses pixel coordinates and homography transformations. This substitution eliminates intermediate processing steps and reduces overall processing time.
Data Source
AI summary
Embodiments of the disclosed technology are directed to using pixel data from two or images of the same real-world region to determine the orientation of substantially planar surfaces in those images. The disclosed technology may be utilized to generate positions and/or orientations of planar surfaces based on multiple 2D images of the planar surface. Implementations can include modeling a planar surface, which can include capturing at least two images (from corresponding different, known positions), determining, from the images, regions that correspond to substantially the same portion of the real-world planar surface, computing a similarity metric for the two regions, and then determining the orientation of the planar surface based on maximizing the similarity metric over various different regions from the images. The described embodiments provide improvements in speed and accuracy compared to existing procedures and are able to find planar surfaces in both low-texture environments and occluded environments.