This article presents a case study that demonstrates the applicability of unmanned aerial vehicle (UAV) photogrammetric data to land surface deformation monitoring in areas affected by underground mining. The results presented include data from two objects located in the Upper Silesian Coal Basin in Poland. The limits of coordinate and displacement accuracy are determined by comparing UAV-derived photogrammetric products to reference data. Vertical displacements are determined based on differences between digital surface models created using UAV imagery from several measurement series. Interpretation problems related to vegetation growth on the terrain surface that significantly affect vertical displacement error are pointed out. Horizontal displacements are determined based on points of observation lines established in the field for monitoring purposes, as well as based on scattered situational details. The use of this type of processing is limited by the need for unambiguous situational details with clear contours. Such details are easy to find in urbanized areas but difficult to find in fields and meadows. In addition, various types of discontinuous deformations are detected and their development over time is presented. The results are compared to forecasted land deformations. As a result of the data processing, it has been estimated that the accuracy of the determination of XY coordinates and the horizontal displacements (RMS) in best case scenario is on the level of 1.5–2 GSD, and about 2–3 GSD for heights and subsidence.

This study presents an approach to the problem of minimizing the impact of low vegetation on the accuracy of a UAV-derived DEM, based on the use of a deep neural network (DNN). It is proposed to use the U-Net network to determine corrections to the height of the raw point cloud so that the processed data reflect the actual earth’s surface. The implemented solution is therefore based on regression, not classification. As a result of the proposed processing method, the expected value of the land surface height is determined for each point of the unified point cloud. In addition, a second U-Net network is trained, enabling the uncertainty of the corrected heights of the land surface to be determined for each point of the unified cloud. The training set includes data from different seasons, which makes the models more resistant and allows for assessment of the impact of the season and more generally the related vegetation status on the model accuracy. The processing results can be used in DEM generation, and also for determining the vertical displacements of the terrain surface associated with underground mining, as well as natural phenomena such as landslides. A key advantage of the proposed processing method is the ability to predict the uncertainty of the results.