Augmented 3D Mapping With Spatial Reference Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Generating accurate three-dimensional maps of real-world environments is computationally difficult and requires significant expenditures in computer vision and robotic equipment.
Innovation Solution
A 3D mapping system that uses annotation of large image data sets with programmatically labeled terrestrial imagery, augmented with remotely sensed overhead image data and location data, implements photogrammetry to create point clouds, and co-registers them with aerial LIDAR data for precise spatial coordinates, stitching together different patches to form a complete 3D map.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If robotic and computer vision systems are implemented to map real-world environments, then mapping capability is improved, but computational complexity and equipment expenditure increase significantly
Solution Approach 1:
The patent segments the 3D mapping process into multiple independent modules: image acquisition from multiple devices, feature point detection, bundle adjustment optimization, and map merging. Each module processes data independently and can be optimized separately, reducing overall computational complexity while maintaining mapping reliability.
Solution Approach 2:
The patent performs preliminary actions by pre-detecting feature points in images and pre-optimizing bundle adjustment parameters before final map generation. This preliminary processing reduces the computational burden during real-time mapping operations and enables more efficient resource utilization.
2Manufacturing precision
If traditional photogrammetry is used to create point clouds, then 3D structure generation is achieved, but accuracy and spatial coordination precision deteriorate
Solution Approach 1:
The patent implements feedback mechanisms through iterative bundle adjustment optimization, where the system continuously refines 3D point coordinates by comparing projected 2D feature points with actual image observations. This feedback loop progressively improves spatial coordinate precision and 3D structure accuracy until convergence is achieved.
Solution Approach 2:
The patent combines multiple data sources including images from terrestrial devices, aerial images, and existing 3D maps to create composite point cloud data. This composite approach integrates information from different perspectives and sources, enhancing both accuracy and spatial coordinate precision through data fusion.
3Area of stationary object
If multiple client device images are processed individually, then comprehensive coverage is achieved, but processing time and computational resources increase
Solution Approach 1:
The patent merges 3D maps generated from multiple client devices by aligning their coordinate systems and integrating feature points. This merging process achieves comprehensive geographic coverage while processing images in parallel batches, maintaining processing efficiency through coordinated multi-device operation rather than sequential processing.
Data Source
AI summary
Systems and method for generating three-dimensional (3D) maps of physical structures are provided. A 3D mapping system can be configured to generate a 3D map of a physical structures using image data sets generated from different user devices. The mapping system can implement reference geometry into a 3D structure generation pipeline to generate more accurate 3D models.