A New and Efficient Method for the Generation of a Ground Control Network with Large Area of Coverage

This paper presents a new and efficient method for the generation of a ground control network that is both accurate and consistent (high relative accuracy) across its entire area of coverage, which can range from a few kilometers squared to the size of countries and conceivably continents. Due to its unique generation process and data content (the 3D locations of the ground control points, their multi-point error covariance, image patches, and ancillary metadata), the network is termed the Metric Information Network (MIN). The MIN is generated and maintained using only the standard inputs and outputs from a series of individual image block adjustments (triangulations) in conjunction with a fusion algorithm. The fusion of information from a new image block adjustment with the current MIN improves the accuracy of all ground control points already in the network and allows for the addition of new points in an optimal and theoretically correct manner. Furthermore, this approach is capable of ingesting and exploiting externally surveyed points, with the fusion of new image block adjustments propagating the surveyed points’ information (accuracy) to the other points in the network. A hypothetical MIN implementation: Use a series of image block adjustments of commercial satellite imagery, along with a sparse set of surveyed ground control points to generate a MIN, and in turn use it to provide control and check points for future image block adjustments, exploitation products, or quality assurance. This paper also presents quantitative MIN performance results related to this scenario.