Underground mining operations result in displacements and deformations of the land surface, which may pose a threat to building structures. The scale of horizontal displacements usually does not exceed several decimeters. These displacements should be monitored to assess and minimize their harmful effects. One of the methods of their observation is aerial photogrammetry. There are solutions that allow for the automation of horizontal displacement determination based on photogrammetric products. In most cases, however, they were created and used to process spaceborne or aerial imagery (with resolutions from about a few decimeters to several meters) and to analyze displacements on a much larger scale. This article proposes a workflow for automatic determination of the field of horizontal displacements caused by underground mining with the use of ultra-high resolution orthomosaics. The study included a comparison of the effectiveness of image registration algorithms for matching of multi-temporal orthomosaics. The outlier removal process is an integral part of the proposed workflow. The results showed that the weighted normalized cross correlation algorithm has the greatest potential for determining displacements based on UAV-derived orthomosaics, while the feature detection and matching algorithms turned out to be less effective in this task. After applying the proposed outlier removal path, the obtained accuracy of determining the displacements is at the level of 1–2 pixels, which was tested for two independent study areas. Accuracy was assessed both in comparison to displacements determined manually on the basis of UAV-derived orthomosaics, and in comparison to displacements independently determined using terrestrial laser scanning.