Using AI-Automated Planetary Computer Cloud Programming for Mapping (Models from Iraq)

Abstract

The research deals with a new experience in using open-source Planetary Computer application programming automated with artificial intelligence and machine learning. It includes a data catalog that relies on a computing environment (Hub) and application programming (STAC) to index and discover geospatial data. The research aims to Learn about some applications of Environmental monitoring capabilities, and explore the data set of satellite images (MODIS, Sentinel-1, viirs,) and choose specific models for application to Iraq to draw objective maps such as (Land surface temperature during the day, MODIS vegetation indices, building footprint using machine learning using DNNs, the highest resolution land cover powered by artificial intelligence, surface water JRC), and in view of the limited research space, the two researchers decided to choose experimental models for the platform to explore the data, tools, and capabilities available in it. Application steps were carried out with an explanation of the programming codes for each map that can visually distinguish spatial variation for different years. The research included three axes: First: the concept of the planetary computer from Microsoft and artificial intelligence applications, second: reading the data catalog from STAC application programming, third: the applied aspect of producing geospatial data maps.

The most important findings of the research are that the cloud platform contains a large geographic database for the spatial environment for different years. The importance of the platform’s tools and the efficiency of the accuracy of the maps became clear through the selected models. The research recommends providing this data through twinning with the specialized company to provide a modern database for Iraq in order to Conducting studies in different regions.