Converting Drone Magnetic Data to Ground Data Using Artificial Intelligence

Mineral exploration necessitates a comprehensive approach that involves analyzing various geophysical, geological, and geochemical datasets, in addition to employing efficient and effective methodologies. Successfully addressing this challenge involves integrating and analyzing diverse geographic data, which often come in different formats and possess distinct features, with the aid of innovative applications. One promising technique involves utilizing artificial intelligence to convert low-resolution drone-collected data into high-resolution ground data. For this particular investigation, three supervised regression models—linear regression, random forest, and enhanced gradient—were implemented in the Python programming environment using magnetometric data obtained from both UAV and Proton ground devices. After evaluating the statistical results, including metrics such as mean square error and mean absolute error, it was determined that the enhanced gradient model outperformed the others. This model exhibited respective values of 0.0004 and 0.01 for training data, 0.001 and 0.02 for experimental data, and 0.001 and 0.01 for validation data. Additionally, the enhanced gradient model demonstrated stability, leading to its selection as the preferred model for prediction purposes.