High‐Accuracy Geological Deformation Monitoring Through Unmanned Aerial Vehicle Image Registration Without Control Points

ABSTRACT Unmanned aerial vehicles (UAVs) have gained popularity in geological deformation monitoring due to their flexibility, cost‐effectiveness, and ease of image acquisition. However, accurate co‐registration of multi‐phase UAV images, essential for effective monitoring, requires numerous ground control points (GCPs), challenging in deformation‐prone areas. This study introduces a UAV‐based monitoring method that eliminates the need for GCPs. Initially, two‐phase images are integrated for joint aerial triangulation. The first‐phase image and tie points undergo bundle adjustment refinement. Control points are identified as tie points with minimal deviation and used to refine the second‐phase image. Three‐dimensional (3D) models are generated from these refined adjustments and differential analysis between the models reveals geological deformation. Experimental results show that this method automatically identifies stable areas and effectively captures both the location and extent of geological deformation, achieving a relative change accuracy of 0.06 m, which is twice the average ground resolution. Compared to traditional aerial triangulation methods without control points, this approach offers more than a tenfold increase in accuracy. The accuracy is on par with methods that utilize multiple control points, indicating the promising potential of this approach for geological deformation monitoring.