The AI for Earth program offered by Microsoft includes an open data program that hosts key geospatial and conservation datasets on the AI for Earth cloud. To enhance the open data catalog, Microsoft’s team is working on processing global atmospherically corrected Landsat and Sentinel imagery archives, for which they were in need of a high-quality global elevation dataset.

However, selecting a quality global elevation dataset can be a challenging task, as there are many options available in the market, each with their own unique features and limitations. One of the main challenges faced by Microsoft was finding a dataset that was accurate, up-to-date, and consistent across different regions of the world. The dataset also had to meet Microsoft’s requirements for resolution, coverage, and format.

Another challenge was to find a dataset that was publicly available or could be licensed at an affordable price. Many commercial datasets are prohibitively expensive for many organizations, including for non-profit organizations like AI for Earth.

To overcome these challenges, Microsoft had to conduct extensive research and evaluation of various datasets. They had to compare the features and limitations of each dataset, and assess their accuracy, completeness, and consistency. They also had to consider the licensing terms and pricing of each dataset, and the level of technical support offered by the vendors.