3D Reconstructability Prediction Using Image-Spatial Feature Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for predicting reconstructability in 3D city reconstruction are inaccurate due to errors in rough scene models, which affect the quality of route planning and subsequent image reconstruction.
Innovation Solution
A method that combines spatial and image characteristics to predict reconstructability by determining spatial relationships between viewpoints and target sampling points, using multilayer perceptrons and transformer models to learn and generate weight matrices for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If reconstructability is calculated using a rough scene model, then the calculation process is simple and fast, but the accuracy of reconstructability prediction is low due to errors in the rough scene model
Solution Approach 1:
The patent introduces a transformer model as an intermediary component that processes the relationship between image characteristics and spatial characteristics to predict reconstructability. The transformer model acts as a mediator that combines information from both sources, resolving the contradiction by adding a specialized processing layer that improves accuracy without requiring complete system redesign
Solution Approach 2:
The patent combines multiple types of characteristics (image characteristics from captured images and spatial characteristics from rough geometric models) to create a composite prediction system. By integrating these different data sources through the transformer model, the system achieves higher accuracy while maintaining manageable complexity through modular architecture
2Measurement precision
If multiple acquisition viewpoints are selected to improve reconstruction quality, then the reconstructability prediction becomes more accurate, but the time required for route planning increases
Solution Approach 1:
The patent performs preliminary extraction of image characteristics and spatial characteristics before the actual route planning. By pre-processing the data and using the transformer model to predict reconstructability in advance, the system reduces the time needed during route planning while maintaining high accuracy in reconstructability assessment
Solution Approach 2:
The patent replaces traditional mechanical/iterative route planning methods with a data-driven transformer model that can quickly assess reconstructability. This substitution allows for faster evaluation by using learned patterns from training data rather than exhaustive computational methods
Data Source
AI summary
Disclosed are a method for predicting reconstructability, a computer device, and a storage medium. In the method, a plurality of viewpoints to be evaluated for a target sampling point are obtained. The target sampling point is located on a rough geometric model. A spatial characteristic of the target sampling point is obtained based on spatial relationships between the plurality of viewpoints to be evaluated and the target sampling point. An image characteristic of the target sampling point is extracted from a target captured image based on a plurality of pre-acquisition viewpoints. The pre-acquisition viewpoints are obtained based on poses of a camera capturing the target captured image. The target captured image is an image containing the target sampling point. The predicting reconstructability for the target sample point is predicted based on the image characteristic and the spatial characteristic.