Artificial Parcel Boundary Mapping From Geospatial Imagery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing geospatial data systems face challenges in generating accurate legal land parcel data due to decentralized availability, high costs, and the indirect nature of parcel boundaries, which are not directly depicted in imagery, necessitating a method to efficiently process mixed coverage areas and convert between vector and raster formats.
Innovation Solution
A machine learning model, utilizing a deep learning architecture, is trained to recognize visual features like fences, roads, and tree lines to demarcate legal land parcels, generating artificial parcel data from geospatial imagery, and converting between vector and raster formats for efficient data processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning models are trained to recognize visual features for parcel boundary detection, then data availability and accuracy are improved in areas without ground truth data, but device complexity and computational resources increase
Solution Approach 1:
The system performs preliminary actions by pre-training machine learning models on areas with ground truth parcel data before deploying them to generate artificial parcel data in areas without ground truth. This prepares the model in advance with learned visual features, enabling it to reliably generate parcel boundaries in target areas without requiring real-time ground truth data.
Solution Approach 2:
The system creates copies of ground truth parcel data by using the machine learning model to generate artificial parcel data that replicates the structure and characteristics of verified parcel data. The model learns from ground truth examples and generates synthetic copies that maintain the statistical and spatial properties of real parcel boundaries, expanding data availability without requiring additional ground truth collection.
2Productivity
If machine learning models process large areas with mixed coverage, then productivity and data generation speed are improved, but measurement precision and accuracy may deteriorate
Solution Approach 1:
The system applies local quality by adapting the machine learning model's behavior based on the local data availability characteristics. In areas with high ground truth coverage, the model is trained with higher precision requirements, while in areas with mixed or no coverage, it generates artificial data with appropriate uncertainty modeling. This allows high productivity across large areas while maintaining measurement precision where ground truth is available for validation.
Solution Approach 2:
The system uses partial action by processing large areas with mixed coverage differently than areas with complete ground truth data. Rather than applying the same rigorous validation everywhere, it generates artificial parcel data in coverage gaps while using ground truth data for validation where available. This partial application of full validation maintains productivity while preserving measurement precision in critical areas.
3Ease of operation
If vector and raster format conversions are implemented, then ease of operation and data processing efficiency are improved, but device complexity increases
Solution Approach 1:
The system implements universality by creating a multi-functional data processing platform that handles both vector and raster formats within a single machine learning framework. The model can ingest training data in either format, process geospatial imagery, and output artificial parcel data in the requested format. This universal capability improves ease of operation by allowing users to work with their preferred format while the system automatically handles conversions as needed.
Data Source
AI summary
Methods and systems for generating artificial parcel data are provided. An example method involves accessing geospatial imagery that depicts at least one building and its surrounding area, applying at least one machine learning model to the geospatial imagery to generate artificial parcel data that represents a shape and location of a legal land parcel occupied by the building, accessing building footprint data comprising at least one polygon that represents a shape and location of the building, storing the artificial parcel data and the building footprint data in association with address data that represents an address of the building, receiving a query, the query comprising an address of the building or geospatial coordinates situated within the building footprint or artificial parcel data corresponding to the building, and providing, in response to the query, the building footprint data and the artificial parcel data corresponding to the building.