Super-Resolution Techniques in Photogrammetric 3D Reconstruction from Close-Range UAV Imagery

Current Multi-View Stereo (MVS) algorithms are tools for high-quality 3D model reconstruction, strongly depending on image spatial resolution. In this context, the combination of Super-Resolution (SR) with image-based reconstruction is turning into an interesting research topic in photogrammetry, around which however only a few works have been reported so far literature. Here, thorough study carried out various state-of-the-art SR techniques to evaluate suitability such approach terms its inclusion process. Deep-learning tested here UAV dataset, while MVS task then performed via Agisoft Metashape photogrammetric tool. The data under experimentation oblique cultural heritage imagery. According results, point clouds from low-resolution images present quality inferior those upsampled high-resolution ones. HAT and DRLN outperform bicubic interpolation, yielding high precision/recall scores differences reconstructed reference surface. current indicates resolution increased by may indeed be advantageous state-of-the art reconstruction.