Pre-Existing 3D Model Updating Through Structural Descriptors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D models of physical structures, particularly buildings, suffer from limited texture resolution, geometry quality, and difficulty in updating, and querying these models for structural modifications yields inaccurate results.
Innovation Solution
A computer-implemented method using trained neural networks to generate descriptors from 2D images of modified structures, comparing these descriptors with pre-existing 3D models to identify and update the models with structural changes, employing techniques like Hamming distance and overlapping metrics for precise matching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional 3D modeling methods using aerial imagery are used, then 3D models can be generated, but the texture resolution and geometry quality are limited
Solution Approach 1:
The patent uses pre-existing 3D models as templates or copies that can be queried and updated. Instead of generating 3D models from scratch using complex aerial imagery processing, the system creates copies or representations of existing models and updates them with new structural information, thereby improving texture resolution and geometry quality without proportionally increasing modeling complexity
Solution Approach 2:
The patent performs preliminary actions by pre-generating 3D models and storing them in a database with associated descriptors. This preliminary modeling work is done in advance, allowing later queries and updates to focus only on detecting structural modifications rather than creating entire models, thus improving quality without proportional increase in complexity
2Measurement precision
If pre-existing 3D models are queried to identify target models, then model retrieval can be performed, but the querying process yields inaccurate results
Solution Approach 1:
The patent segments the 3D model representation into distinct structural features and generates separate descriptors for each feature (e.g., windows, doors, roof type). This segmentation allows the querying system to compare specific structural elements independently, significantly improving query accuracy by focusing on key discriminative features rather than attempting to match entire complex models at once
Solution Approach 2:
The patent transforms 3D model data into descriptor parameters that capture essential structural characteristics. By changing the representation from raw geometric data to structured descriptors with specific parameters, the system enables more accurate comparison and detection of structural modifications, improving query accuracy while managing detection complexity
3Reliability
If 3D models are updated to reflect structural modifications, then model currency is improved, but the updating process is difficult and time-consuming
Solution Approach 1:
The patent extracts only the essential structural features and their descriptors from the 3D models that are necessary for identification and update purposes. By taking out only the critical geometric and textural parameters rather than updating entire complex model structures, the system improves model currency efficiently by focusing updates on extracted key features
Solution Approach 2:
The patent performs preliminary actions by pre-computing and storing descriptors for all possible structural features in the database. This preliminary preparation allows update operations to simply compare new descriptors against stored ones and apply changes only where differences are detected, significantly improving update efficiency while maintaining model currency
Data Source
AI summary
Identifying a pre-existing three-dimensional (3D) model of a target structure includes receiving at least one two-dimensional (2D) image of a target physical structure; generating a predicted 3D model of the target structure based on the at least one 2D image; generating a search descriptor of the predicted 3D model; querying a data structure storing a plurality of pre-existing descriptors, where each pre-existing descriptor characterizes a previously constructed 3D model of an associated physical structure; and identifying at least one previously constructed 3D model that is substantially similar to the predicted 3D model based on a difference between the search descriptor and the plurality of pre-existing descriptors.