Lidar point clouds (LPC), elevation models, GPS data, image mosaics, and aerial images from thermal infra-red (TIR) and natural color (RGB) cameras collected during UAS operations at Lower Darby Creek, Darby Township, Pennsylvania, March 13 to 17, 2024

RTK-GPS Base Station and Rover: A Spectra Precision model SP80 GNSS base station receiving real-time differential corrections from the satellites was set up to collect positions in Receiver Independent Exchange (RINEX) format at 1 hz on a known point prior to and for the duration of lidar data surveys. The base station was operating every day of field collection between 3/13/2024 - 3/17/2024. A SP80 rover receiving real-time differential corrections from the Pennsylvania CORS network was used to survey points across the field area, including ground position validation points on 3/17/2024. The RTK receivers for both the base and rover are linked via Bluetooth to a handheld data collector (Carlson CHC LT30 Handheld Terminal running Carlson SurvCE v. 4.06 software under Windows Mobile v. 6.1 Professional operating system). Rover antennas were mounted on 2 m survey rods with bubble levels and 5 cm (2 inch) sand feet. Conversions from satellite coordinates to NAD83(2011) UTM zone 18N (EPSG::6347) and NAVD88 (EPSG::5703) were made by Carlson SurvCE software in the data collector.

Ground control: AeroPoint targets were emplaced on flat, stable surfaces, with clear view of the sky and left undisturbed, collecting data for at least 2 hours. AeroPoint data was uploaded via a WiFi connection to a user defined account on propelleraero.com. The raw data are processed by Propeller and exported to a CSV file in NAD83(2011) UTM zone 18N (EPSG::6347) for horizontal and NAVD88 (geoid 18) (EPSG::5703) for vertical.

Lidar System: A YellowScan Mapper Plus (YSMP) high-density lidar system containing a lidar sensor and natural color (RGB) camera module was mounted to a DJI Matrice 600 (M600) uncrewed aircraft system (UAS) to collect low-altitude aerial lidar point cloud (LPC) data and simultaneous overlapping natural color (RGB) images. Lidar flights occurred between 3/13/2024 and 3/17/2024. A configuration text file (CONFIG.TXT) is pre-loaded onto the lidar USB thumb drive and was edited to control lidar and camera module settings, including the camera triggering height, the camera triggering interval, and the lidar scan pattern. For this field area over marsh and forested terrain, a point density of ~300 points per square meter (ppsqm) was targeted with a non-repetitive (spirograph) scan pattern. The flight plans for this project targeted a swath overlap of 50% with the following flight parameters: altitude of 61 m, flight line spacing at 45 m, a velocity of 10 m/s, and a 2-second camera triggering interval. At these parameters, the natural color RGB camera module was set to trigger every 3 seconds to achieve 50% sidelap and forelap for colorization of the LPC in post-processing. For every lidar UAS flight, the following steps were followed to properly operate the system and collect data: After powering on the UAS and lidar system, the GPS, scanner, and camera module all must initialize. After takeoff, an additional initialization pattern must be manually flown by the pilot to signal to the IMU to begin collecting lidar data. This pattern consists of a forward, back, and forward flight path ending with a hook-shaped coordinated turn on the last forward maneuver. It is best practice that this is flown below the camera triggering altitude. Once the maneuver is completed, the UAS is ascended to the camera triggering altitude where the RGB camera module sets the ISO for the entire flight. It is recommended to make this altitude close to the flight plan altitude. The mission is uploaded to the UAS and it begins the automated flight plan. When the data collection is finished or batteries require swapping, the same initialization pattern flown at the beginning is flown again to de-initialize the scanner. The lidar data is saved to a 256 GB USB thumb drive in three different files: (1) the IMU+GPS data, decimated for quick post-processing, in binary *.ys format, (2) the scanner data in *.lvx format (~600 MB per minute of data collection), (3) and the complete IMU+GPS data in Applanix (Trimble) binary *.t04 format. The RGB images are saved to a 64 GB micro-SD card as *.jpeg files. The camera triggering time-interval was determined and set based on the focal length and ground sampling distance (GSD) of the camera as well as the altitude and speed of the UAS in order to achieve ~50% forelap between adjacent photos for colorizing the lidar point cloud. The camera triggering interval as well as the altitude at which the camera automatically sets the ISO for the scene is written into a configuration file located on a USB stick that is plugged into the YellowScan Mapper Plus system.

Thermal Camera: Thermal infra-red (TIR) and natural color (RGB) aerial images were collected using a DJI Zenmuse XT2 camera mounted to a DJI M600 UAS. TIR and RGB imagery flights occurred between 3/13/2024 and 3/16/2024. The XT2 camera was set to high gain scene range, which has a higher sensitivity to temperature differences over a smaller range (-13 to 212 degrees Fahrenheit). The Darby Creek main channel and tributaries were surveyed with corridor type mapping flight plan design. Flight plans targeted a photo sidelap and forelap of 75%, sufficient for photogrammetric processing by following these flight parameters: 61 meters altitude, 3 m/s velocity, and a 3-second camera triggering interval. The TIR and RGB images were saved to a 128 GB micro-SD card as *.tiff and *.jpeg files, respectively.

Lidar data processing: After each data collection, the three data files are stored on the USB thumb drive in a folder following a naming scheme of "YS-YYYYMMDD-HHMMSS" where YS is the acronym for YellowScan, YYYYMMDD is the date stamp, and HHMMSS is the time stamp in UTC time (e.g. YS-20230830-171402). Within this folder are the three files containing the IMU+GPS data in .ys and .t04 format and the scanner data in *.lvx format, which follow the same naming scheme as the parent folder. The RGB images are stored on the SD card following a naming scheme of DSC#####.JPG (e.g. DSC00001.JPG). Upon import of the raw scanner data (.ys file) for each flight to YellowScan CloudStation v. 2210.0.0, the sensor's Lidar (.profile) and Camera (.camera) profiles, provided by the vendor, are selected for the project and the project coordinate reference system is defined. CloudStation auto-identifies sensor position trajectories that are to be included in the model based on their length and linearity, but adjustments are usually required as the software will sometimes misidentify trajectories that should not be included, such as the path the UAS took to reach the first waypoint of the mission or leave out paths that are considered too short. If there is a slight waiver in a flight path, the software will sometimes cut off the ends of the flight paths. The .T04 file and base station GNSS RINEX file were used to correct and optimize the sensor position trajectories using th GNSS Inertial Processor and produce a Smoothed Best Estimate of Trajectory (SBET) file in .txt ASCII format, which represents the Post Processing Kinematic (PPK) Solution. The lever arm offsets and boresight angle corrections were applied to the LPC, along with a strip adjustment between transect swaths using CloudStation's "robust" setting. The LPC was then colorized using the images. The Cut Overlap function was applied to remove redundant points based on the closest parallel flight trajectory. The unclassified LPC model was then exported from CloudStation as a 1.4 LAZ file. The LPCs were brought into Global Mapper (v. 26.0) and manually cleaned by reclassifying or removing points interpreted to be noise, points that were either not classified or incorrectly classified as ground or are above a reasonable elevation threshold based on the surrounding features. The LPCs were exported as a 1.4 LAZ files. To produce the DSMs, the LPCs are gridded in Global Mapper using maximum values and exported to geotiff format. To produce DTMs, only the ground points are gridded using maximum values and exported to geotiff format. All files were exported in EPSG:6347 NAD83(2011)/UTM zone 18N and the vertical datum in EPSG:5703 NAVD88 using Geoid 18.

Imagery processing: Telemetry navigation logs for each flight were exported from the Ignis app in GPX format and used to manually geotag the imagery from the YellowScan Mapper Plus camera modules in preparation for photogrammetric processing. Image timestamps were adjusted to UTC time depending on time zone and daylight saving. These corrections were done using the program ExifTool. ExifTool uses linear interpolation to determine the position of the device at the time recorded in the image. This creates values for tags GPSLatitude, GPS LatitudeRef, GPSLongitude, GPSLongitudeRef, GPSAltitude, GPSTimeStamp, GPSDateStamp. A sample of code used to geotag images: exiftool -v2 -geotag ../../telemetry/Nav_FlightLogs/Flight\ Logs\ GPX/flight_log_2022_10_10_14_02_42_AGL.gpx -geosync=+16 '-geotime<${DateTimeOriginal}-04:00' -P f05XT2c/*.JPG. All images were renamed to ensure unique filenames following the USGS Coastal and Marine Hazards and Resources Program's best practices for image naming convention: "FieldActivityNumber_f##sensorID_YYYYMMDDTHHMMSS_OriginalFilename.tif" (e.g. "2024006FA_f05XT2c_20240313T155702Z_DJI_0184.jpg") Field activity number is a unique identifier assigned to the field effort under which the data was collected (e.g. 2024006FA), f## denotes the flight in which the photos were taken (e.g. f05), sensor ID is the sensor or camera used to collect the imagery (e.g. XT2c = Zenmuse XT2 natural color RGB camera; XT2t = Zenmuse XT2 thermal infra-red TIR camera, YSMP = YellowScan Mapper Plus), 'YYYYMMDD' is the UTC date, 'HHMMSS' is UTC time and the 'T' is used to separate the UTC date from UTC time, and the original filename is appended to the end. To ensure compliance with USGS Coastal and Marine Hazards and Resources Program's best practices for image exif, exif tags were added using a CSV containing header information for each photo. A sample of the code used to edit the image tags using ExifTool version 12.52:
'exiftool -csv="C:\directory\name\EXIF.csv" C:\directory\name\of\photos *.[ext]
The following tags were added to the image exif: -IPTC:Credit="U.S. Geological Survey" -IPTC:Contact="WHSC_data_contact@usgs.gov" -EXIF:Copyright="Public Domain" -XMP:UsageTerms="Unless otherwise stated, all data, metadata and related materials are considered to satisfy the quality standards relative to the purpose for which the data were collected. Although these data and associated metadata have been reviewed for accuracy and completeness and approved for release by the U.S. Geological Survey (USGS), no warranty expressed or implied is made regarding the display or utility of the data for other purposes, nor on all computer systems, nor shall the act of distribution constitute any such warranty." -EXIF:ImageDescription="https://cmgds.marine.usgs.gov/fan_info.php?fan=2024-006-FA; Low-altitude aerial natural color photograph of the Lower Darby Creek Superfund Site and John Heinz National Wildlife Refuge, Darby Township, Pennsylvania, taken with a DJI XT2 thermal camera during USGS field activity 2024-006-FA" -XMP:AttributionURL="https://doi.org/10.5066/P134HU3Y" -EXIF:GPSAreaInformation="camera-integrated GPS" -EXIF:GPSMapDatum="EPSG:4326 (WGS 84)" -IPTC:keywords="UAS Aerial Imagery, John Heinz National Wildlife Refuge, Lower Darby Creek Superfund, Darby Township, Pennsylvania, 2023-006-FA, USGS, DJI Zenmuse XT2" -EXIF:Artist="WHCMSC AIM Group" -XMP:PreservedFilename>Filename.
The images for this data release have been uploaded to the Imagery Data System (IDS), which renames the images to safeguard from accidental overwriting. A user can restore the preserved filename with ExifTool. It is recommended to test on a single image before running on all images. To test on a single image:
exiftool "-XMP:PreservedFilename<Filename" filename.tif.
Once desired output is confirmed, use this code to rename all images in the folder: exiftool "-Filename<XMP:PreservedFilename" -overwrite_original -P *.tif *note that the case of the image extension may matter (e.g. TIF vs tif).
CSV files corresponding to each sensor set of photos containing image location information was generated using ExifTool version 12.52. A sample of this code and the included tags:exiftool -n -csv -XMP:PreservedFilename -EXIF:GPSMapDatum -EXIF:GPSLatitude -EXIF:GPSLatitudeRef -EXIF:GPSLongitude -EXIF:GPSLongitudeRef *.JPG > 2023020FA_YSM+_ImageLocs.csv

Structure-from-motion processing: The SfM products were created in Agisoft Metashape v. 2.0.1 using the following generalized workflow (see Over and others, 2021 for a more detailed methodology explanation): 1. A project was created in Metashape and imagery specific to each sensor and flight area were imported. 2. Photos were aligned at a low accuracy and then GCPs were automatically detected in the point cloud. GCP positions (2024006FA_Clearview_Mar2024_AeroPoint.csv) were added to the project in the reference systems NAD83(2011)/UTM Zone 18N and NAVD88 (geoid 18). The horizontal and vertical accuracies for the GCPs were set to 0.04/0.02 m, respectively, and the camera positions for the images were turned off. The photos were then re-aligned with high accuracy (the pixels were not subsampled) using a key point limit of 60,000 and unlimited tie points. 3. The alignment process matched pixels between images to create point clouds and put the imagery into a relative spatial context using the GCPs. The resultant point clouds were filtered using one iteration of the 'Reconstruction uncertainty' filter at a level of 10, one iteration of the 'Projection accuracy' filter at a level of 3, and eight iterations of the 'Reprojection accuracy' filter to get to a level of 0.3. With each filter, iteration points are selected, deleted, and then the camera model was optimized to refine the focal length, cx, cy, k1, k2, k3, p1, and p2 camera model coefficients. 4. Multiple ‘chunks’ were created so that a high-quality dense cloud with a low-frequency filtering algorithm could be made from the images. The dense point cloud was then edited by visual inspection and Metashape’s confidence filter to remove points with a low confidence near the edges and near water bodies. 5. A DSM was built from the dense point cloud and then an orthomosaic was built from the DSM with refined seamlines. The DSMs are exported in the coordinate reference system (CRS) EPSG:6347 NAD83(2011)/UTM18N coordinate reference system (CRS) and the vertical reference system EPSG:5703 NAVD88 (Geoid 18).

CLOUD OPTIMIZATION: All geoTIFF products were DEFLATE compressed and turned into a cloud-optimized GeoTIFFs (COG) using either the export tool in Global Mapper or gdal_translate with the following command: for %i in (.\*.tif) do gdal_translate %i .\cog\%~ni_cog.tif -of COG -stats -co BLOCKSIZE=256 -co COMPRESS=DEFLATE -co PREDICTOR=YES -co NUM_THREADS=ALL_CPUS -co BIGTIFF=YES (v. 3.1.4 accessed October 20, 2020 https://gdal.org/). Where "i" is the name of each geoTIFF section.