AI Height Seed Initialization for Digital Elevation Model Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current geospatial modeling systems face challenges in efficiently generating accurate digital surface models and change detection, particularly with large-scale satellite data of inhomogeneous quality and resolution, often requiring expert-driven manual processes and struggling with image registry and processing complexity.
Innovation Solution
An AI system utilizing a convolutional neural network (CNN) for estimating height maps from electro-optic imagery, incorporating semantic segmentation and game theory optimization to improve pixel height estimation and generate digital surface models, while also enabling efficient change detection through deep learning and semantic change detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional geospatial modeling systems use manual expert-driven processes for generating digital surface models, then measurement precision may be maintained, but productivity is significantly reduced
Solution Approach 1:
The patent replaces manual expert-driven mechanical processes with an automated AI system using convolutional neural networks. The CNN model automatically estimates pixel heights from satellite imagery, eliminating the need for manual expert intervention while maintaining accuracy through trained algorithms and semantic segmentation techniques.
Solution Approach 2:
The system enables self-service automation where the AI model independently processes satellite imagery to generate digital surface models without requiring continuous human expert involvement. The trained CNN performs autonomous height estimation, allowing the system to serve itself in processing large-scale geospatial data efficiently.
2Area of stationary object
If traditional systems process large-scale satellite data with inhomogeneous quality and resolution, then comprehensive coverage is achieved, but device complexity increases
Solution Approach 1:
The patent transforms the complexity of processing inhomogeneous satellite data by changing the approach from traditional manual processing to AI-based automated processing. The CNN model adapts to varying data quality and resolution through trained parameters, automatically adjusting to handle inhomogeneous inputs without requiring complex manual intervention protocols.
Solution Approach 2:
The AI system provides universal processing capability that handles multiple types of satellite imagery with varying quality and resolution through a single trained model. The convolutional neural network is designed to process diverse geospatial data formats and conditions uniformly, reducing the need for multiple specialized processing systems.
3Measurement precision
If manual expert processes are used for change detection, then detection accuracy is maintained, but loss of time increases
Solution Approach 1:
The patent replaces manual expert-driven change detection processes with automated AI-based analysis. The convolutional neural network automatically identifies and detects changes in geospatial data by comparing different time periods, eliminating the need for manual expert review while maintaining detection accuracy through learned patterns and semantic segmentation.
Data Source
AI summary
An artificial intelligence (AI) system for generating a digital surface model (DSM) may include a memory and a processor cooperating therewith to determine an estimated height map from electro-optic (EO) imagery of a geographic area using artificial intelligence. The processor may further generate cost coefficients for a three-dimensional (3D) cost cube based upon stereo-geographic image data and height value seeding using the estimated height map, and generate a DSM for the geographic area based upon the 3D cost cube and outputting the DSM to a display.