UAS Orthomosaic Imagery and Lidar-Derived Digital Elevation Models for the Cape May, Railroad, and Ocean City Confined Disposal Facilities, New Jersey, USA

These UAS lidar and imagery (red, green, blue bands) data were collected as part of work funded by the US Army Corps of Engineers Dredging Operations Environmental Research (Work Unit: 25-02 Enhancing Confined Disposal Facilities Operation to Support Coastal Resiliency), the USACE Philadelphia District, and the National Coastal Mapping Program. These data are useful for characterizing the volumes of dredged materials stored in these confined dredged material disposal facilities as well as assessing the dikes and shorelines surrounding these dredged material management areas.

A general metadata document is included as well as accuracy reports.

Given the purpose of these datasets for applied engineering in the United States, the projected coordinate system for the datasets is NAD 1983 State Plane New Jersey FIPS 2900 (US Feet) and the vertical coordinate system in the datasets is the North American Vertical Datum of 1988 (US survey feet). These can easily be converted for other purposes in most standard Geographic Information System software. In some instances, there is a good amount of glare on the water and seamlines from the different images in the orthomosaics, which were not the primary data need and instead are supplementary to the lidar data.  Low tide and time windows of collection dictated the collection more than better imagery conditions. 

Ground Control Data

For each site, ground control data was collected using two Leica GS18 Real-Time Kinematic (RTK) Global Navigation Satellite System (GNSS) equipment. A survey point was created out of a 20 cm length of rebar. A static base station was established over this point with a collection interval of one (1) second. A total collection time greater than three (3) hours was achieved for each session. To establish ground truth data, points were collected dispersed across the area of interest, however, due to vegetation some areas were not surveyed for ground control data. Ground control points (GCPs), points that are surveyed to control and/or compared to the lidar data, were collected using a Leica GS18 RTK rover. Points were identified as photo id (PID), non-vegetated vertical accuracy (NVA), and vegetated vertical accuracy (VVA) points. GCPs that were collected on photo id targets, were surveyed on the center of a “checkerboard” marker.  All NVAs were collected on available hard flat surfaces, and VVAs were collected on the ground beneath vegetation to check lidar penetration. All points were collected and post-processed in North American Datum (NAD) of 1983 with the 2011 realization and ellipsoid heights tied to Geodetic Reference System (GRS) 1980.  Rinex files from the base station data were submitted to Online User Positioning Service (OPUS) for the computation of a more refined solution of the static base station coordinates. All GCPs are adjusted utilizing this refined solution for the static base station in Leica Infinity software. Overall, the RTK GNSS consistently maintained an accuracy below (2) cm Root Mean Square Error (RMSE).   

Equipment and Flight Information

Lidar surveys in 2024 were conducted utilizing a UAS mounted GeoCue True View 640 carried on an Inspired Flight 1200. The True View 640 utilizes Riegel miniVUX-3UAV sensor, Applanix APX-20 GNSS with dual Inertial Measurement Units (IMU) and dual Sony 1” CMOS IMX-183 RGB cameras. The cameras are mounted at 25 degrees cross track off nadir. The Riegl miniVUX-3UAV sensor is a 360-degree scanner capable of five (5) returns per pulse producing 200,000 points per second over a 120-degree field of view. Lidar surveys in 2025 were conducted utilizing a UAS mounted GeoCue True View 660. The 660 uses triple Sony 1” CMOS IMX-183 RGB cameras where two (2) mounted at 25 degrees cross track off nadir, and one (1) camera is mounted nadir.

The UAS was assembled and flown over the Cape May Campground Confined Disposal Facility (CDF) site on the April 16th, 2024, and July 20th, 2025. The 2024 survey was primarily for the CDF area whereas the 2025 survey included the shoreline along the Cape May Canal. One (1) flight was flown during the 2024 data collection. The flight altitude was 60 m above ground level (AGL), and 60 m line to line spacing. During the 2025 survey, there was seven (7) flights conducted for the Campground CDF. Flightlines oriented North and South were flown in four (4) flights at 60 m AGL and 60 m line spacing. Due to dense vegetation, two (2) flights were flown over the CDF. One (1) was around the outer containment berm to increase chances of penetrating vegetation and detecting bottom. This flight was at 60 m AGL and 30 m line to line spacing. Another flight was flown with flightlines oriented East and West direction at 40 m AGL and 40 m line spacing over the CDF. Finally, a flight was flown during low tide along the shoreline. This flight was flown at 60 m AGL and 40 m spacing. 

Flights were conducted on April 18th, 2024 over the Railroad CDF. Two (2) flights were performed flying Northwest and Southeast orientation over this site. Both were at 60 m AGL and 60 m line to line spacing.

Flights were conducted on July 24, 2025 over Ocean City CDF. Flights were planned to collect data at low tide to increase data collection area of the topographic lidar sensor. Two (2) flights were conducted flying North and South. These flights were planned at 60 m AGL and 60 m line to line spacing. A third flight was performed over the CDF to increase point density and vegetation penetration. The altitude was 40m and 40m line spacing.

For all flights, the speed was set to five (5) m/s during collection. With these settings, the lidar system will have a 40% or greater overlap of lines and single pass average point density of 113 pts/m². Images are collected simultaneously with lidar data and are triggered every two seconds. 

Lidar Processing and DEM Creation

At this time a statistical comparison of the GCPs and lidar point cloud was performed (outlined in more detail later in this section). This ensures good alignment before creating final data products. The .las files were imported into to QT Modeler for further quality control checks and digital elevation model (DEM) creation. There the point cloud was filtered to show only ground classified points. A quality control check was performed to detect missed noise or misclassed ground points in the data. Once the check was complete, a 50 cm grid digital terrain model (DTM) was created with mean-z adaptive triangulation. This DTM was imported into LP360 where it was used to process the imagery orthomosaics. LP360 uses a built in Agisoft Metashape tool for processing of imagery. Images are selected during lidar trajectory processing as to which to retain based on computed flight lines. This eliminates images in turns and unselected flight lines, such as transition lines. The retained photos are used in orthomosaic processing. The DTM that was generated for orthomosaic processing is imported into LP360. The DTM is selected to be used for orthomosaic generation. Alignment accruacy is set to “High”, and the Ground Sampling Distance (GSD) is set to “Optimal from Images.” After processing is completed, a manual quality control check is performed to ensure streaking, blurring, and seamline visibility are at an acceptable amount. The ground control is then compared to the orthomosaic and will be discussed in further detail later in this report. Imagery data collected in the 2025 data collection was processed using Coorelator3d software. The orthomosaic is then reprojected into NAD83 State Plane New Jersey and NAVD88 US Survey Foot in ArcPro.

Ground control points were imported and processed in Leica Infinity software. These points were exported as a .csv file and imported into LP360 for comparisons with the UAS lidar data. Once imported into LP360, the “checkerboard” PIDs’ centers were manually selected from the UAS imagery to provide the horizontal comparisons. Only PIDs were used for horizontal comparisons and planimetric errors were computed from these. The GCPs were compared to the ground classified lidar points for vertical accuracies. The LP360 software computed statistical errors on the vertical and planimetric comparisons. For the Campground CDF collection in 2024, the reported NVA RMSEz is 1.2 cm and vertical mean error was 0.3 cm. The PID planimetric RMSE is 4.5 cm and mean error is 4.0 cm. For the Railroad 2024 collection only PIDs and VVAs were collected. The RMSEz for the PIDs is 2.2 cm and vertical mean error is 1.5 cm. The planimetric RMSE from the PIDs is 3.7 cm and planimetric mean error is 3.2 cm. The VVA RMSEz is 10.8 cm and vertical mean error is –9.2 cm. The Campground CDF 2025 the NVA RMSEz is 4.2cm with a vertical mean error of 0.3 cm. Planimetric RMSE was 5.1 cm and planimetric mean error was 4.8 cm. VVA points were collected in the 2025 survey and were used for vertical comparisons to the lidar data. The VVA RMSEz was 13.5 cm and mean error was -8.2 cm. Ocean City CDF NVA RMSEz was 2.1 cm with a mean error of 0.0 cm. The planimetric RMSE is 6.6 cm and the planimetric mean errors is 6.2 cm. The VVE RMSEz is 10.2 cm and a vertical mean error of -6.7 cm.

The point cloud was then converted, using Vdatum, to NAD83 State Plane New Jersey and North American Vertical Datum of 1988 (NAVD88) both in US Survey Foot. The State Plane point cloud was then used to create subsequent DEMs. A DTM was created from the ground classified .las files in QT Modeler using a one (1) foot grid. The DTM was created from the ground points using adaptive triangulation method and a mean-z algorithm and was unrestricted on interpolation distance. A digital surface model (DSM) was created using all points from classes one (1) and two (2). The DSM used a 1-foot grid and utilizing an adaptive triangulation method and a max-z algorithm. Interpolation distance was limited to three (3) foot distance to a real point and nine (9) foot max triangular side. The DTM is then clipped in ArcPro using the data from the DSM as the clipping device. This allows the DTM be interpolated across sections where data was captured, but missing ground points as in beneath vegetation or buildings, but limits interpolation across areas where no data was captured, such as water or edges of flight path. 

Imagery Processing and Orthomosaic Creation

These files contain 3-band true-color RGB, orthorectified mosaic imagery.

Camera data were stored as Joint Photographic Experts Group (JPEG) files created during the data collection.  The UAS on-board navigation provided planned GPS exposure points commensurate with lidar collection.  Positioning data (Latitude, Longitude, and Altitude; roll, pitch, and heading) was stored as EXIF data in the JPEG as well as in separate external orientation (EO) Comma Separated (.csv) file.

The unrectified JPEG files, navigation EO data and the DEM file as well as camera calibration parameters were given as input to Simactive's Correlator 3D (C3D).  C3D is a software package for processing lidar, true color imagery and mulitspectral imagery. Using C3D, Aerial Triangulation (AT) was performed using unconstrained optimization for the intrinsic parameters and RTK/PPK (Real-Time Kinematic/Post-Processed Kinematic) Assisted optimization for the exterior orientation parameters since the position and orientation information for each frame was refined during post processing of the trajectory.  Ground Control Points (GCPs) were only used for accuracy determinations during post-mosaic QA/QC.  The lidar derived DTM was used to further constrain the data to true ground. The image data was geometrically corrected and orthorectifed using the DSM.  From there, a mosaic was created from the orthorectified images. The mosaic data were exported from C3D as a GEOTIF (*.tif) 16-bit file in a single block accordingly as collected.

Disclaimer:

These data provide elevation values and are only accurate for the time associated with the DEMs. Acknowledgement of the authors would be appreciated in any publications or derived products. The data herein, including but not limited to geographic data, tabular data, analytical data, electronic data structures or files, are provided 'as is' without warranty of any kind, either expressed or implied, or statutory, including, but not limited to, the implied warranties or merchantability and fitness for a particular purpose. The entire risk as to the quality and performance of the data is assumed by the user. No guarantee of accuracy is granted, nor is any responsibility for reliance thereon assumed. In no event shall the U.S. Army Corps of Engineers, the ERDC, or the JALBTCX, be liable for direct, indirect, incidental, consequential or special damages of any kind, including, but not limited to, loss of anticipated profits or benefits arising out of use of or reliance on the data. The U.S. Army Corps of Engineers, the ERDC, and the JALBTCX do not accept liability for any damages or misrepresentation caused by inaccuracies in the data or as a result of changes to the data caused by system transfers or other transformations or conversions, nor is there responsibility assumed to maintain the data in any manner or form.

File Naming Notes:

DEM = Digital Elevation Model

SPC = State Plane Coordinates

Ft = US Survey Feet

NAVD88 = North American Vertical Datum of 1988

DTM = Digital Terrain Model

Ortho = orthomosaic red-green-blue band imagery

vva = vegetated vertical accuracy

nva = non-vegetated vertical accuracy

pid = photo identification points

groundshot = all ground points regardless of vegetation presence/absences