In the globalized economic world, it’s become important to understand the purpose behind infrastructural initiatives occurring within undeveloped regions of the earth, especially when the financing for such projects must be coming from external sources.  Global coverage from a large number of commercial, private, and government satellites have produced enormous imagery datasets. Thus, these geospatial resources can be easily mined and processed using machine learning algorithms and neural networks. Increasingly beneficial is the fact that this data is available in largely undeveloped areas where ground or aerial coverage is either non-existent or not commonly acquired, such as major portions of the African continent. Although the advantages of such easily accessible, large datasets are substantial, the downside is that a majority of these geospatial data resources are in a state of technical static, as it’s difficult to quickly understand the importance of a particular image from the moment in time it was taken. This work seeks to leverage portions of the Author’s previous work presented at I/ITSEC 2019, for fully automated data segmentation and object information extraction framework for creating simulation terrain using UAV-based photogrammetric data (Chen et al. 2019) by extending some of the feature classification methodology to satellite imagery, with a focus on change detection in large-scale human construction and development, such as railroads.  This research hopes to evaluate specific events over time that are easily and rapidly detectable. While we will utilize existing architecture from our current methods, a new set of training data will be produced from satellite imagery for detecting this infrastructure. A goal of this research is to allow automated monitoring of change over time for large-scale infrastructure projects to best determine reliable metrics that define the scope and scale, as well as the future direction these construction initiatives could be expected to take.