USGS CoastCam CACO-01 at Head of the Meadow Beach, Cape Cod National Seashore, Massachusetts (2019 to 2024): Calibration Data

Before installation at Head of the Meadow CACO-01 c1 and c2 were used to take a series of photos for the IO calibration. This process consisted of collecting 86 and 100 JPEG images for c1 and c2, respectively (see CACO01_CalibrationImageList.csv and the CACO01_IO collection on the IDS). These images contain a 28 mm square checkerboard moved throughout the image frame. This checkerboard pattern allows lens distortion metrics to be calculated. The image file name is formatted as CACO01_c#_IO_YYYYMMDD_######.jpg where c# is the camera number, YYYYMMDD is the date the IO images were obtained, and ###### is the original number written to the file. Note that the time and location of where the actual photos were taken is irrelevant, they have been tagged with the GPS location of the camera pole for which they are relevant and given a GPSTimeStamp of 12:00:00. Step processed by Jonathan Borden.

To place the photos taken by the CoastCam into real world coordinates and to rectify them into a planar view, objects within the FOV of the camera must be located. To do so, the elevation and position of the camera lenses c1 and c2 in the CoastCam setup were measured with the RTK GPS and then multiple stationary targets, between 5 and 10 per camera as recommended by Brodie and Bruder (2020), were spread through the camera FOV to be used as GCPs, and the location of each was measured with the RTK GPS equipment as follows.
1. A base station was set up in the parking lot at Head of the Meadow Beach on an established benchmark.
2. A GPS receiver was connected to the base station to collect RTK data in NAD83(2011)/ UTM Zone 19N + NAVD88 using GEOID 12B* on March 4, 2020 and in NAD83(2011)/ UTM Zone 19N + NAVD88 using GEOID 18 on January 26, 2022.
3. Five plywood checkerboard targets were placed on the beach propped up facing towards CoastCam CACO-01. These were measured with the rovers at the intersection of the sand and the middle of the target. Once all GCPs were placed, a snapshot (snap) was taken with the camera. The targets were moved to multiple locations to take additional snap images.
On 20200304, four snap images were taken with c1 (CACO01_c1_20200304_GCP_image_raw_1583337600.jpg, CACO01_c1_20200304_GCP_image_raw_1583341200.jpg, CACO01_c1_20200304_GCP_image_raw_1583344800.jpg, and CACO01_c1_20200304_GCP_image_raw_1583346600.jpg) for a total of 10 unique GCP targets. All of the GCPs can be viewed on CACO01_c1_20200304_GCP_image_combined_raw.jpg where targets were cropped and superimposed onto the image, they are annotated in CACO01_c1_20200304_GCP_image_annotated.jpg to match the GCP# in CACO01_c1_20200304_EO_GCPsurvey.csv. The image coordinates of the targets on the combined image are the same as in the original images.
On 20200304, three snap images were taken with c2 (CACO01_c2_20200304_GCP_image_raw_1583337600.jpg, CACO01_c2_20200304_GCP_image_raw_1583341200.jpg, and CACO01_c1_20200304_GCP_image_raw_1583344800.jpg) for a total of 6 unique GCP targets. Note there are four targets in *raw_1583341200 but only the target in the foreground was surveyed with the RTK. All of the GCPs can be viewed on CACO01_c2_20200304_GCP_image_combined_raw.jpg where useable targets were cropped and superimposed onto the image, they are annotated in CACO01_c2_20200304_GCP_image_annotated.jpg to match the GCP# in CACO01_c2_20200304_EO_GCPsurvey.csv. The image coordinates of the targets on the combined image are the same as in the original images.
On 20220126, one snap image was taken with c1 (CACO01_c1_20220126_GCP_image_raw.jpg) for a total of 5 unique GCP targets. The targets are annotated in CACO01_c1_20220126_GCP_image_annotated.jpg to match the GCP# in CACO01_c1_20220126_EO_GCPsurvey.csv.
On 20220126, one snap image was taken with c2 (CACO01_c2_20220126_GCP_image_raw.jpg) for a total of 5 unique GCP targets. The targets are annotated in CACO01_c2_20220126_GCP_image_annotated.jpg to match the GCP# in CACO01_c2_20220126_EO_GCPsurvey.csv.
The image file name is formatted as CACO01_c#_YYYYMMDD_GCP_type_######.jpg where c# is the camera number, YYYYMMDD is the date the EO images were obtained, type is raw, combined or annotated, and ###### is the original number written to the file. *The elevation values using GEOID 12B were transformed to GEOID 18 values using vdatum (https://vdatum.noaa.gov/vdatumweb/) on 2025-02-01 and no changes in the values occurred, so for consistency all data is provided in GEOID 18.

The IO images together with the GCP targets from the EO surveys were used in the CIRN Toolbox and with the Camera Calibration App (MATLAB) to solve for the 3 radial coefficients and 2 tangential coefficients that make up the intrinsic parameters and lens distortion and the EO parameters. The IO coefficient values are provided in YAML files for each camera (CACO01_c1_20191212_IO.yaml and CACO01_c2_20191212_IO.yaml) and the EO parameters are provided in YAML files for each camera (CACO01_c1_20200304_EO.yaml, CACO01_c2_20200304_EO.yaml, CACO01_c1_20220126_EO.yaml, and CACO01_c2_20220126_EO.yaml). Files are provided in zip folder by date and combine camera c1 and c2 files; CACO01_IO_20191212.zip, CACO01_EO_20200304.zip, and CACO01_EO_20220126.zip.
These YAML files can be re-solved for using the raw images and survey data provided in the CIRN Toolbox. Please refer to Brodie and Bruder (2020) for information about the methods. Step processed by Athina Lange.

All images were geotagged and processed to add additional information required by the USGS to the exchangeable image file format (EXIF) and other imagery headers using PyExifTool Python library (https://pypi.org/project/PyExifTool/, version: 0.5.6). A list of all the useable images is provided in CACO01_ImageList.csv.
The following describes an example of the Python dictionary object with each entry in the dictionary being a tag-value pair written to each image in the data release: tags = { 'Copyright': 'Public Domain', 'CopyrightNotice': 'Public Domain - please credit U.S. Geological Survey', 'Artist': 'USGS WHCMSC', 'Make': 'FLIR', 'Model': 'Blackfly 50S5C-C', 'GPSLatitude': '42.05049347, 'GPSLongitude': '-70.07738362', 'GPSMapDatum': 'EPSG:6318 NAD83 (2011)', 'GPSAreaInformation': 'camera pole location surveyed with RTK-GNSS', 'GPSLatitudeRef': 'N', 'GPSLongitudeRef': 'W', 'Credit': 'U.S. Geological Survey', 'Contact': 'whsc_data_contact@usgs.gov', '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', 'AttributionURL': 'https://doi.org/10.5066/P95C0DGM', 'Author': 'Jin-Si R. Over, Christopher R. Sherwood, Eric D. Marsjanik, Jonathan S. Borden, Athina M.Z. Lange', 'XMP-photoshop:Credit': 'U.S. Geological Survey', 'iptcCore:CreatorWorkEmail': 'whsc_data_contact@usgs.gov','Rights': 'Public Domain', 'ExternalMetadataLink': 'https://www1.usgs.gov/pir/api/identifiers/USGS:60d61173d34e960393675c90'}.
The UserComment EXIF metadata field was used to store the camera intrinsics and extrinsics in the image metadata. The calibration date refers to the date the nearest EO was done (not relevant). The UserComment has a nested dictionary structure. The highest-level dictionary contains two nested dictionaries as well as a note explaining the structure of the UserComment field. One nested dictionary contains the relevant intrinsic (NU, NV, coU, coV, fx, fy, d1, d2, d3, t1, t2) and extrinsic (x, y, z, azimuth, tilt, roll) data variables. The extrinsic variables include both horizontal and vertical position of the camera in UTM coordinates. The other nested dictionary contains text descriptions for each of the variables in the dictionary of data variables. The GPSLatitude and GPSLongitude are specified as the CACO-01 camera pole position and allows the user to find all relevant imagery for CACO-02 on the Imagery Data Service.
Additionally, tags were added to the Python dictionary object that have different values for either an IO or EO calibration image, the FA relevant to the extrinsics, and camera (c#) c1 or c2. These tags are included as follows:
IO image: 'Description':'https://cmgds.marine.usgs.gov/services/activity.php?fan=2019-043-FA; This image was captured from a stationary camera (CACO-01 [c#]) to perform an intrinsic calibration.'
'Keywords': "Head of the Meadow Beach, Cape Cod National Seashore, Massachusetts, CoastCam, USGS, intrinsic calibration, remote sensing"
EO image: 'Description':'https://cmgds.marine.usgs.gov/services/activity.php?fan=[INSERT FA]; This image was captured from a stationary camera (CACO-01 [c#]) at Cape Cod National Seashore, Massachusetts with extrinsic calibration targets in the field of view.'
'Keywords': "Head of the Meadow Beach, Cape Cod National Seashore, Massachusetts, CoastCam, USGS, extrinsic calibration, remote sensing, ground control point"
To read out the imagery header information to a CSV file the command is: 'exiftool -csv *.jpg > allheaders.csv'. Specific tags may be specified with this command, if preferred. Or use the online tool https://exif.tools/ (accessed May 9, 2025)

The calibration images were uploaded to the IDS into the collections CACO01_EO and CACO01_IO. These can be accessed at https://cmgds.marine.usgs.gov/idsviewer/data_release/10.5066-P95C0DGM.