Geospatial Modeling System for DSM Building and Vegetation Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for separating building and vegetation data in digital surface models (DSMs), particularly those generated stereographically, are time-consuming and labor-intensive, often requiring manual annotation due to the expense and complexity of LIDAR data collection.
Innovation Solution
A geospatial modeling system that includes a database and processor to classify points in DSMs based on spectral range differences, using multi-spectral image data to automatically separate bare earth data from building and vegetation, and classify pixels through a voting process, enabling efficient separation of building and vegetation data even in stereographic DSMs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If LIDAR data is used to separate building and vegetation data in DSMs, then classification accuracy is improved, but data collection cost and time increase significantly
Solution Approach 1:
The patent uses multi-spectral image data as a substitute (copy) for expensive LIDAR data. By processing readily available optical imagery with spectral information, the system achieves building and vegetation separation without requiring costly LIDAR data collection, thus maintaining classification accuracy while eliminating the time and cost penalty of LIDAR acquisition
Solution Approach 2:
The patent transforms the classification approach by utilizing spectral range parameters from multi-spectral images. Instead of relying on LIDAR's direct distance measurement capability, the system uses spectral reflectance characteristics across multiple bands to differentiate buildings from vegetation, achieving accurate classification through parameter substitution
2Measurement precision
If manual annotation is used to separate building and vegetation data in stereographic DSMs, then classification accuracy is improved, but labor intensity and time consumption increase
Solution Approach 1:
The patent replaces the manual mechanical annotation process with an automated computational system. The processor automatically analyzes multi-spectral image data and applies classification algorithms to separate buildings from vegetation in stereographic DSMs, eliminating manual labor while maintaining or improving classification accuracy through consistent algorithmic application
Solution Approach 2:
The system enables self-service classification where the processing algorithm automatically identifies and separates building and vegetation features without human intervention. The multi-spectral data and classification algorithms work together autonomously to produce classified DSMs, freeing users from time-consuming manual annotation tasks
3Productivity
If automated classification is used for building and vegetation separation, then productivity is improved, but measurement precision may deteriorate
Solution Approach 1:
The patent enhances automated classification precision by utilizing multiple spectral bands rather than single-band or grayscale data. The multi-spectral parameters provide richer information about material properties, enabling the automated system to distinguish buildings from vegetation more accurately while maintaining high processing efficiency through algorithmic automation
Solution Approach 2:
The system uses multi-spectral image data as a information-rich substitute that enables automated classification to achieve accuracy comparable to or exceeding manual methods. The spectral information acts as an enhanced data copy that provides additional discriminative features for automated algorithms to achieve high precision without manual intervention
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system automates the separation of building and vegetation data in DSMs, reducing manual effort and costs, and effectively processes stereographic DSMs with improved accuracy and efficiency compared to traditional methods.
Implementation Method 1
at least one spectral range indicative of a difference between buildings and vegetation
Data Source
AI summary
A geospatial modeling system may include a geospatial model database configured to store a digital surface model (DSM) of a geographical area, and to store image data of the geographical area. The image data may have a spectral range indicative of a difference between buildings and vegetation. The geospatial modeling system may also include a processor cooperating with the geospatial model database to separate bare earth data from remaining building and vegetation data in the DSM to define a building and vegetation DSM. The processor may also register the image data with the building and vegetation DSM, and classify each point of the building and vegetation DSM as either building or vegetation based upon the spectral range of the image data.