Orthomosaic and digital surface model of the main Casey station buildings, 12th February 2021.

Full description

Images were acquired from approximately 80 m above ground surface on the 12th of February 2021, using a Phantom 4 Advanced drone with an FC330 camera. The images are in file input_images.zip.

The mission planning software DJI GS Pro was used to automatically acquire images at suitable locations across the survey area to enable the reconstruction of a three dimensional model.

Images 422 to 531 were imported to the photogrammetry software Pix4D (version 4.6.4). The created Pix4D project is Station12Feb2021_limited.p4d, and the processing report is Station12Feb2021_limited_report.pdf.

Four three-dimensional ground control points were used to improve the positioning of the model. No two dimensional control points or check points were used.

These points were in ITRF 2000@2000 datum (UTM Zone 49S), with co-ordinates as per the table below:

Label, Type, X(m), Y(m), Z(m), Accuracy Horz(m), Accuracy Vert(M)
BM05, 3D GCP, 478814.460, 2648561.910, 38.558, 0.050, 0.100
EW-05, 3D GCP, 478635.540, 2648617.260, 27.260, 0.050, 0.100
FuelFlange, 3D GCP, 478970.810, 2648642.250, 21.920, 0.050, 0.100
MeltbellFootingA, 3D GCP, 478680.270, 2648466.547, 35.850, 0.050, 0.100

BM-05 is a survey benchmark near the Casey flagpoles, see https://data.aad.gov.au/aadc/survey/display_station.cfm?station_id=600
EW-05 is a 44 gallon drum used as a groundwater extraction well by the remediation project
Fuel Flange is the last fuel flange located on the elevated fuel line prior to the fuel line “dipping” under the wharf road.
Meltbell footing A is a concrete footing for the Casey melt bell (surveyed in 2019/20).

No point cloud processing (e.g. removal of errant points) was done prior to orthomosaic and model generation.

The resulting orthomosaic (Station12Feb2021_limited_transparent_mosaic_group1.tif) has an average ground sampling distance of 2.9 cm, and covers an area of approximately 15.8 hectares, encompassing the majority of buildings along “main street” at Casey. The quarry, biopiles, helipad, and upper fuel farm area are all visible.

Contour lines were generated in Pix4D at 0.5 m intervals.

Due to the limited number of ground control points, and their imprecision, the estimated residual mean squared error across three dimensions is 0.17 m (17cm), and will be worse on the periphery of the imaged area.

The orthomosaic was exported from ArcGIS to a Google Earth file (CaseyStation Orthomosaic Feb 12 2021.kmz) using XTools Pro Version 17.2.

A map was created in ArcGIS showing the orthomosaic with a background showing contour lines obtained from the AADC data product windmill_is.mdb.

The map was exported in .jpg and .pdf format at 250 dpi.
Casey Station Orthomosaic Feb 12 2021.pdf
Casey Station Orthomosaic Feb 12 2021.jpg

The Pix4D folder structure has been copied across (with the exception of the temp folder) and is included in this dataset.

Pix4D Folder Structure:

Station12Feb2021_limited.zip
1_intitial
• Contains Pix4D files created during the project
• Contains the final processing report (as .pdf)
2_densification
• Contains the 3D mesh as an .obj file
• Contains the point cloud as a .LAS and .PLY file
• Contains the point cloud as a .p4b file
3_dsm_ortho
• Contains the digital surface model as a georeferenced .tif file
• Contains the orthomosaic as a georeferenced .tif file

A text readable log file from the project processing is in the file Station12Feb2021_limited.log

Lineage

Progress Code: completed

Statement: No ground control points were set out or surveyed for this project, so ground control points had to be located from the photos and previously surveyed points. Due to the height of the drone and thus the resolution of the images, it was difficult to precisely locate the ground control points in the photographs used to tie the model. In particular, it was difficult to locate BM05 (a survey benchmark defined by a small brass plaque). The other three ground control points also suffered from small locational errors, with the largest errors likely expected to be in the z dimension. The processing report provides a guide to the project accuracy (approximately 20cm in the X,Y and Z dimensions). However, this orthomosaic, point cloud and contours should not be used for accurate geolocation.