Phase II year
2019
(last award dollars: 1709924435)
NGA seeks a solution to provide highly accurate foundation data of buildings, similar features, and linear structures by conflating information from multiple images and contributed content. Our solution AMCON Geo (Automated Multi-source Conflation, Geospatial) is a framework for combining multiple sources together to improve overall relative and absolute accuracies of the data, and increase operational throughput by putting all of the data together in an easy to visualize, and easy to distribute, data package. The resulting conflation contains estimates of goodness of fit for the mapping and propagates underlying estimates of the quality of the source data via the following: - Data sources are converted into the appropriate foundation components; - Beginning with the 3D geometry layers, matching geometry features are automatically found for each data source;- Perform a non-rigid, geometry-preserving mapping to each data layer, bringing its features into alignment with the underlying model; - Use mutual information to perform a global alignment, minimizing the mapping errors across all datasets; - Create a geodatabase from the aggregated data; and - Visualize the results in an exploratory 3D viewer. We believe AMCON Geo aligns well with our proven expertise and breadth of ongoing, overlapping DoD initiatives.
