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本数据集包含无人机系统（Unmanned Aircraft System, UAS）激光雷达（lidar）与红绿蓝波段影像数据，其采集工作由美国陆军工程兵团（United States Army Corps of Engineers, USACE）疏浚作业环境研究项目（工作单元：25-02 强化密闭处置设施运行以支撑海岸韧性）、USACE费城分局以及国家海岸测绘计划资助。本数据集可用于表征密闭疏浚物料处置设施内存储的疏浚物料体积，同时可用于评估疏浚物料管理区域周边的堤坝与岸线。 本数据集附带通用元数据文档与精度报告。 鉴于本数据集面向美国应用工程场景，其投影坐标系采用1983北美基准州平面新泽西州分区（FIPS 2900，美国测量英尺），垂直坐标系采用1988北美垂直基准面（美国测量英尺）。多数标准地理信息系统软件均可轻松完成坐标系转换。部分正射镶嵌影像中存在水体眩光与不同影像间的拼接缝，此类影像并非核心数据，仅作为激光雷达数据的补充。数据采集受低潮时段与作业窗口限制，优先保障采集时效性而非最优成像条件。 ## 地面控制数据 针对每个采集站点，均采用两台徕卡（Leica）GS18实时动态（Real-Time Kinematic, RTK）全球导航卫星系统（Global Navigation Satellite System, GNSS）设备采集地面控制数据。地面控制点采用20厘米长的钢筋制作，以此为基准搭建静态基准站，采集间隔设为1秒，单次采集时长均超过3小时。为获取地面真值数据，在研究区域内分散布设采集点，但受植被遮挡影响，部分区域未完成地面控制数据采集。 地面控制点（Ground Control Point, GCP）采用徕卡GS18 RTK流动站采集，可分为照片识别点（Photo ID, PID）、非植被区垂直精度点（Non-Vegetated Vertical Accuracy, NVA）与植被区垂直精度点（Vegetated Vertical Accuracy, VVA）三类。其中，照片识别点需在“棋盘格”标记的中心位置采集；非植被区垂直精度点均采集于裸露硬质平面；植被区垂直精度点则设于植被下方地面，用于验证激光雷达的植被穿透能力。所有点位均采用1983北美基准（NAD 1983，2011实现版）进行后处理，椭球高度绑定1980大地参考系统（Geodetic Reference System, GRS 1980）。基准站采集的Rinex文件提交至在线用户定位服务（Online User Positioning Service, OPUS），以优化静态基准站坐标的解算结果。所有地面控制点均基于该优化后的基准站坐标，通过徕卡Leica Infinity软件进行平差处理。整体而言，RTK GNSS的测量精度稳定控制在2厘米均方根误差（Root Mean Square Error, RMSE）以内。 ## 设备与飞行信息 2024年的激光雷达测绘采用搭载于Inspired Flight 1200无人机的GeoCue True View 640系统完成。该系统集成Riegl miniVUX-3UAV传感器、Applanix APX-20 GNSS模块（搭载双惯性测量单元（Inertial Measurement Unit, IMU））以及双台索尼1英寸CMOS IMX-183 RGB相机，相机安装角度为偏离天底点横滚25度。Riegl miniVUX-3UAV传感器为360度扫描器，支持每个脉冲返回5个点云，在120度视场角下可实现每秒20万个点的采样率。 2025年的激光雷达测绘采用GeoCue True View 660系统，该系统搭载三台索尼1英寸CMOS IMX-183 RGB相机：两台安装于偏离天底点横滚25度位置，另一台正对天底安装。 测绘无人机分别于2024年4月16日、2025年7月20日飞越开普梅露营地密闭处置设施（Confined Disposal Facility, CDF）站点。2024年测绘仅覆盖密闭处置设施区域，仅执行1次飞行任务，飞行高度为地面以上60米（Above Ground Level, AGL），航线间距为60米。2025年测绘新增开普梅运河沿岸岸线区域，共执行7次飞行任务：其中4次采用南北向航线，飞行高度60米AGL、航线间距60米；针对植被茂密区域，执行2次飞行作业：一次沿外围围堤飞行，以提升植被穿透与底部探测概率，飞行高度60米AGL、航线间距30米；另一次采用东西向航线，飞行高度40米AGL、航线间距40米；最后一次在低潮时段沿岸线飞行，飞行高度60米AGL、航线间距40米。 2024年4月18日，测绘团队飞越铁路密闭处置设施站点，执行2次西北-东南向飞行任务，飞行高度均为60米AGL、航线间距60米。 2025年7月24日，测绘团队完成海洋城密闭处置设施站点的测绘工作。本次测绘计划在低潮时段开展，以提升地形激光雷达传感器的有效采集范围：共执行2次南北向飞行任务，飞行高度60米AGL、航线间距60米；另执行1次飞行任务以提升点云密度与植被穿透能力，飞行高度40米AGL、航线间距40米。 所有飞行任务的飞行速度均设为5米/秒。在此参数设置下，激光雷达系统的航线重叠率可达40%及以上，单次采样平均点密度为113点/平方米。影像数据与激光雷达数据同步采集，触发间隔为2秒。 ## 激光雷达处理与数字高程模型（Digital Elevation Model, DEM）创建 本阶段首先开展地面控制点与激光雷达点云的统计比对（详见本章节后续内容），以确保在生成最终数据产品前实现良好对齐。将.las格式文件导入QT Modeler软件，开展进一步质量控制检查与数字高程模型（DEM）创建工作：对原始点云进行滤波，仅保留地面分类点；执行质量控制检查，以识别数据中遗漏的噪声点或误分类的地面点。检查完成后，采用均值-自适应三角剖分法生成50厘米格网的数字地形模型（Digital Terrain Model, DTM）。将该DTM导入LP360软件，用于影像正射镶嵌处理。LP360内置Agisoft Metashape工具用于影像处理，在激光雷达轨迹处理阶段，基于解算的航线筛选保留可用影像，剔除转弯时段与非作业航线（如过渡航线）对应的影像，将筛选后的影像用于正射镶嵌处理。将用于正射镶嵌的DTM导入LP360，并设置用于正射镶嵌生成，对齐精度设为“高”，地面采样距离（Ground Sampling Distance, GSD）设为“基于影像自动优化”。处理完成后，执行人工质量控制检查，确保影像的条纹、模糊与拼接缝可见度处于可接受范围。随后将地面控制点与正射镶嵌影像进行比对，相关细节将在本报告后续章节展开。2025年采集的影像数据采用Coorelator3d软件进行处理，最终将正射镶嵌影像重投影至NAD83州平面新泽西分区与1988北美垂直基准面坐标系，该操作通过ArcPro软件完成。 地面控制点通过Leica Infinity软件导入并处理，导出为.csv（逗号分隔值）格式文件后，导入LP360软件以与无人机激光雷达数据开展比对。导入LP360后，手动从无人机影像中选取“棋盘格”照片识别点的中心位置，用于平面精度比对，仅采用照片识别点开展平面误差计算。将地面控制点与地面分类激光雷达点云进行比对，以评估垂直精度。LP360软件可自动计算垂直与平面比对的统计误差。 - 2024年开普梅露营地密闭处置设施采集数据：非植被区垂直精度点的RMSEz为1.2厘米，垂直平均误差为0.3厘米；照片识别点的平面RMSE为4.5厘米，平面平均误差为4.0厘米。 - 2024年铁路密闭处置设施采集数据：仅采集照片识别点与植被区垂直精度点。其中照片识别点的RMSEz为2.2厘米，垂直平均误差为1.5厘米；平面RMSE为3.7厘米，平面平均误差为3.2厘米。植被区垂直精度点的RMSEz为10.8厘米，垂直平均误差为-9.2厘米。 - 2025年开普梅露营地密闭处置设施采集数据：非植被区垂直精度点的RMSEz为4.2厘米，垂直平均误差为0.3厘米；平面RMSE为5.1厘米，平面平均误差为4.8厘米。本次采集同步获取植被区垂直精度点，用于与激光雷达数据开展垂直精度比对，其RMSEz为13.5厘米，垂直平均误差为-8.2厘米。 - 2025年海洋城密闭处置设施采集数据：非植被区垂直精度点的RMSEz为2.1厘米，垂直平均误差为0.0厘米；平面RMSE为6.6厘米，平面平均误差为6.2厘米。植被区垂直精度点的RMSEz为10.2厘米，垂直平均误差为-6.7厘米。 将点云文件通过Vdatum软件转换为NAD83州平面新泽西分区与1988北美垂直基准面坐标系，基于转换后的州平面点云生成后续DEM产品。通过QT Modeler软件，从地面分类.las文件中生成1英尺格网的数字地形模型：采用自适应三角剖分法与均值-z算法，基于地面点生成DTM，且不对插值距离设限。采用所有1类与2类点云数据生成数字表面模型（Digital Surface Model, DSM），DSM采用1英尺格网，通过自适应三角剖分法与最大值-z算法生成，插值距离限制为：与真实点的最大距离为3英尺，三角形最大边长为9英尺。随后在ArcPro软件中，以DSM数据为裁剪边界对DTM进行裁剪，该操作可使DTM在有数据采集但缺失地面点的区域（如植被下方或建筑物下方）进行插值，同时限制在无数据采集区域（如水体或航线边缘）的插值行为。 ## 影像处理与正射镶嵌创建 本数据集包含3波段真彩色RGB正射镶嵌影像。 相机采集的影像以联合图像专家组（Joint Photographic Experts Group, JPEG）格式存储。无人机机载导航系统可提供与激光雷达采集同步的GPS曝光点位信息。影像的定位数据（纬度、经度、高度；滚转、俯仰与航向）以EXIF（可交换图像文件格式）数据形式存储于JPEG文件中，同时也保存于独立的外部定向（External Orientation, EO）逗号分隔值（Comma Separated Values, CSV）文件中。 将未校正的JPEG影像、外部定向数据、DEM文件与相机校准参数作为输入，导入Simactive公司的Correlator 3D（C3D）软件。C3D是一款用于处理激光雷达、真彩色影像与多光谱影像的软件包。基于C3D软件，采用无约束优化算法对影像内参开展空中三角测量（Aerial Triangulation, AT），并采用实时动态/后处理动态（Real-Time Kinematic/Post-Processed Kinematic, RTK/PPK）辅助优化算法对影像外参进行解算，因为每帧影像的位置与姿态信息可通过轨迹后处理进行优化。地面控制点仅用于正射镶嵌完成后的精度验证与质量控制。采用激光雷达生成的DTM进一步约束影像，使其贴合真实地面。利用DSM对影像进行几何校正与正射校正，随后基于正射校正后的影像生成镶嵌图。将镶嵌数据以单块地理标记图像文件格式（GeoTIFF, *.tif）16位文件形式从C3D软件导出。 ## 免责声明 本数据集提供的高程值仅对应数字高程模型生成时的时间节点。若将本数据集用于出版物或衍生产品，敬请致谢本数据集的研发团队。 本数据集所包含的所有数据（包括但不限于地理数据、表格数据、分析数据、电子数据结构或文件）均按“现状”提供，不附带任何明示、暗示或法定的担保，包括但不限于适销性与特定用途适用性的默示担保。数据的质量与性能风险由使用者自行承担。本数据集不保证任何精度，也不承担因依赖本数据集而产生的任何责任。在任何情况下，美国陆军工程兵团、工程研发中心（Engineer Research and Development Center, ERDC）与JALBTCX均不对因使用或依赖本数据集而产生的直接、间接、附带、特殊或后果性损害承担责任，包括但不限于预期利润或收益的损失。美国陆军工程兵团、工程研发中心与JALBTCX均不对因数据不准确、系统传输或其他转换操作导致的数据变更所引发的任何损害或误述承担责任，也不承担以任何方式或形式维护本数据集的义务。 ## 文件命名说明 - DEM：数字高程模型（Digital Elevation Model） - SPC：州平面坐标系（State Plane Coordinates） - Ft：美国测量英尺（US Survey Feet） - NAVD88：1988北美垂直基准面（North American Vertical Datum of 1988） - DTM：数字地形模型（Digital Terrain Model） - Ortho：红绿蓝波段正射镶嵌影像 - vva：植被区垂直精度点（Vegetated Vertical Accuracy） - nva：非植被区垂直精度点（Non-Vegetated Vertical Accuracy） - pid：照片识别点（Photo Identification Points） - groundshot：无论植被覆盖与否的所有地面点