SfM Experiment using aerial photos circa 2002
Background
My current research and surveying project involves a small cemetery in the midst of a 43-acre (17.4-hectare) subdivision. The development, which began in the 2000s, had only contour lines depicted on the original survey plan. As a result of the site work done at that time, and in particular as part of the stormwater drainage system and road construction, the bank on 2 sides of the 50’ x 50’ (15.24m x 15.24m) cemetery requires stabilization as soon as possible before we start seeing these early settlers’ and revolutionary war veterans’ remains peeking out from the exposed and continually eroding bank.
In May 2002, the Maine Department of Transportation hired James W. Sewall to fly much of area between Bangor and Trenton, including along the Union River and subject property in Ellsworth at a photo scale of 1:12500. The scanned prints are thankfully archived and available from Digital Maine Explore Maine Geology: Maine Geological Survey: Maine ACF . Whether or not the fiducials included in the scanned images help or hinder RealityScan remains undetermined as there is a scanning glitch seen in the upper right corner of every image and one image; i.e., DOT02-21-14, has its fiducials cropped the along the bottom margin. I’m also uncertain if the scanned images are even of adequate quality to attempt this experiment.
Many thanks to Paul Smitherman of UMO Folger Special Collections, for his help finding the related USGS OSL_2750 Camera Calibration Report. The calibration had been completed about 11 months prior to the flight date. No evidence was found, neither in the archived records, nor on the face of the scanned images, that any ground control was run at the time of the flight in May 2002. Accordingly, synthesized GCPs were created predicated on 1) horizontal data, orthoRegional2020 + 2) vertical data, DEMterrain2021 at selected centerline intersections. The USGS OSL_2750 calibration report for the Zeiss RMK Top 15 camera and Pleogon A3/4 Lens assembly indicates the focal length to be 153.327 mm.
Workflow in RealityScan v2.1.1 on Windows 10 Pro
Began by loading images and GCPs, and getting them tied, but am flailing about getting the Zeiss camera details correctly entered, how to deal (if at all) with the fiducials. Alignments have been sad. Any advise will be very much appreciated! Thank you.
Kind regards,
Kelly
Shared Folder - SfM experiment using 2002 archived imagery
DOT02 information received from Paul Smitherman.zip (13.9 MB)
2002.zip (1.5 GB)
epsg6480 Name,N,E,Ht, meter - synthGCPs.csv (877 bytes)