Since the transition from military to civilian applications, unmanned aerial vehicles have been quickly becoming a platform for spatial data acquisition and providing researchers flexibility at their work, while being affordable and easy to use. In the master thesis we present the operation and historical evolution of unmanned aerial vehicles as well as fields and legal frameworks of their use. We selected three different study areas for which stereo imagery was captured from different flight altitudes in order to determine the impact of the sensor remoteness on the surface model quality. Study areas also differed in the representation of the land cover types and ground surface topology. In addition to the data acquisition, we also presented data processing by the generation of a georeferenced photogrammetric point cloud, which further served to derive a continuous surface model whose quality was also numerically and descriptively evaluated. The model accuracy evaluation was based on comparison with the LiDAR surface model. The comparative analysis also showed whether the type of land cover affects the model accuracy and assessed the suitability as well as pros and cons of such data acquisition for geographical research.