After visual pre-selection, approximate coordinates wereassigned to the photographs based on the synchronised GPS flightpath. Geotagging4software was used to write the coordinates toimage JPEG EXIF headers. Agisoft Photoscan Professional (0.85)software was used for the 3D reconstruction of the camera pos-itions and terrain features. The specific algorithms implemented inPhotoscan are not detailed in the manual, however, a descriptionof the SfM procedure in Photoscan and commonly used parametersare described in Verhoeven (2011). Photoscan follows a commonSfM and multiview stereopsis (MVS) workflow starting with imagefeature identification and feature matching. The approximate GPScoordinates of the camera stations were used at this stage to guidethe matching process. Image matching was carried out with thePhotoScan accuracy set to high. The initial bundle adjustmentresulted in a position and orientation for each camera exposurestation and the 3D coordinates of all image features. These coordi-nates formed a sparse 3D point cloud of the terrain (1.85 millionpoints). A dense geometry reconstruction based on multi-viewstereopsis resulted in a more detailed 3D model with 10 millionfacets. This model was used to identify the 12 large GCPs to improvethe absolute accuracy of the bundle adjustment and the 3D model.After the recomputed bundle adjustment a new 3D model with50 million facets was generated with the PhotoScan accuracy forthe geometry build set to high. Finally, a DSM was generated bygridding the 3D model based on a given cartographic projection(WGS84 UTM 49S) and cell size (2 cm). Projecting the original pho-tographs onto the 3D surface and blending their overlap zonesproduced an orthophoto mosaic (orthomosaic) of the whole area.
Due to the relatively low flying height we were able to generate aDSM at 2 cm resolution and an orthomosaic at 1 cm.An assessment of the geometric accuracy in easting (X), northing(Y), and height (Z) was carried out for the orthomosaic and DSM.Thirty of the small orange GCPs (excluded from georeferencing inthe SfM process) were identified in the orthomosaic. The coordi-nates of the disk centroids were retrieved from the image mosaicand compared to the corresponding surveyed GPS coordinates,resulting in mean and root mean squared error (RMSE) accuracymeasures in the X and Y direction. The height value was derivedfrom the DSM for the GCP centroids and also compared to the GPSobservations, producing mean and RMSE accuracy measures for theZ direction (Höhle and Höhle, 2009; Harwin and Lucieer, 2012).