Image Sharpening Effects on Photogrammetric Product Quality

Richard Massaro

Richard Massaro USACE-ERDC-GRL

15G

Image and video acquisition from unmanned aerial sensors (UAS) has become commonplace. Together with advances in Structure-from-Motion (SfM) algorithms and computational resources, photogrammetry has witnessed a rebirth in recent years. UAS-based image and video acquisition is typically automated and governed by the GPS location and attitude of the platform. Ideally, camera settings and UAS collection parameters are optimized prior to collection so as to provide the sharpest, clearest imagery for photogrammetric processing. However, problems such as wind gusts, focus issues, or UAS platform stabilization may often arise and can blur the imagery. In turn, the blurred imagery may degrade the quality of automatically chosen tie points when the imagery is submitted to a photogrammetric workflow. Meanwhile, there are a number of existing digital image sharpening routines that can be used to rectify image blur. In this study, we evaluate the photogrammetric product quality and accuracy from several small UAS datasets that have been subjected to image blur and those same datasets that have been digitally sharpened. The products include point clouds, orthomosaics, and digital surface models. We also compare the results to ground truth measurements. If a very high level of fidelity and accuracy is required for geospatial products and if blur is suspected in the collected dataset, we suggest submitting the original imagery and/or video collection to a sharpening algorithm prior to photogrammetric processing.

15:45 Image Sharpening Effects on Photogrammetric Product Quality, Richard Massaro

January 29 @ 15:45
15:45 — 15:50 (5′)

Quartz AB

Richard Massaro

Add to Google Calendar