National Pipeline Mapping System: https://pvnpms.phmsa.dot.gov/PublicViewer/TC Energy PDF Map: https://www.tcenergy.com/siteassets/pdfs/natural-gas/gtnxp/tce-gas-transmission-northwest-xpress-map.pdfCompressor data HIFLD (https://ft.maps.arcgis.com/home/item.html?id=d910e5aca7434d19899b1e5a05234051)USGS Topo Maps: https://ngmdb.usgs.gov/topoview/viewer/#4/40.00/-100.00Aerial Imagery:Historical - Google Earth Pro (using the time slider to check for ground scars over the years)Bing Satellite Imagery QGIS Plugin

Summary

Photovoltaic (PV) energy generation plays a crucial role in the energy transition. Small-scale PV installations are deployed at an unprecedented pace, and their integration into the grid can be challenging since stakeholders often lack quality data about these installations. Overhead imagery is increasingly used to improve the knowledge of distributed PV installations with machine learning models capable of automatically mapping these installations. However, these models cannot be easily transferred from one region or data source to another due to differences in image acquisition. To address this issue known as domain shift and foster the development of PV array mapping pipelines, we propose a dataset containing aerial images, annotations, and segmentation masks. We provide installation metadata for more than 28,000 installations. We provide ground truth segmentation masks for 13,000 installations, including 7,000 with annotations for two different image providers. Finally, we provide ground truth annotations and associated installation metadata for more than 8,000 installations. Dataset applications include end-to-end PV registry construction, robust PV installations mapping, and analysis of crowdsourced datasets.

This dataset contains the complete records associated with the article "A crowdsourced dataset of aerial images of solar panels, their segmentation masks, and characteristics", currently under review. The preprint is accessible at this link: https://arxiv.org/abs/2209.03726. These complete records consist of:

The complete training dataset containing RGB overhead imagery, segmentation masks and metadata of PV installations (folder bdappv),

The raw crowdsourcing data, and the postprocessed data for replication and validation (folder data).

Data records

Folders are organized as follows:

bdappv/ Root data folder

google / ign: One folder for each campaign

img/: Folder containing all the images presented to the users. This folder contains 28807 images for Google and 17325 images for IGN.

mask/: Folder containing all segmentations masks generated from the polygon annotations of the users. This folder contains 13303 masks for Google and 7686 masks for IGN.

metadata.csv The .csv file with the installations' metadata.

data/ Root data folder

raw/ Folder containing the raw crowdsourcing data and raw metadata;

input-google.json: .json input data data containing all information on images and raw annotators’ contributions for both phases (clicks and polygons) during the first annotation campaign;

input-ign.json: .json input data containing all information on images and raw annotators’ contributions for both phases (clicks and polygons) during the second annotation campaign;

raw-metadata.json: .json output containing the PV systems’ metadata extracted from the BDPV database before filtering. It can be used to replicate the association between the installations and the segmentation masks, as done in the notebook metadata.

replication/ Folder containing the compiled data used to generate the segmentation masks;

campaign-google/campaign-ign: One folder for each campaign

click-analysis.json: .json output on the click analysis, compiling raw input into a few best-guess locations for the PV arrays. This dataset enables the replication of our annotations,

polygon-analysis.json: .json output of polygon analysis, compiling raw input into a best-guess polygon for the PV arrays.

validation/ Folder containing the compiled data used for technical validation.

campaign-google/campaign-ign: One folder for each campaign

click-analysis-thres=1.0.json: .json output of the click analysis with a lowered threshold to analyze the effect of the threshold on image classification, as done in the notebook annotation;

polygon-analysis-thres=1.0.json: .json output of polygon analysis, with a lowered threshold to analyze the effect of the threshold on polygon annotation, as done in the notebook annotations.

metadata.csv: the .csv file of filtered installations' metadata.

License

We extracted the thumbnails contained in the google/img/ folder using Google Earth Engine API and we generated the thumbnails contained in the ign/img/ folder from high resolution tiles downloaded from the online IGN portal accessible here: https://geoservices.ign.fr/bdortho. Images provided by Google are subjet to Google's terms and conditions. Images provided by the IGN are subject to an open license 2.0.

Access the terms and conditions of Google images at this URL: https://www.google.com/intl/en/help/legalnotices_maps/

Access the terms and conditions of IGN images at this URL: https://www.etalab.gouv.fr/wp-content/uploads/2018/11/open-licence.pdf

This dataset is a comprehensive inventory of Alaskan buildings, storage tanks, and roads that were: (1) detected from 0.5 meter resolution satellite imagery of communities (acquired between 2018-2023) and (2) supplemented by OpenStreetMap data. We created HABITAT (High-resolution Arctic Built Infrastructure and Terrain Analysis Tool), a deep learning-based, high-performance computing-enabled mapping pipeline to automatically detect buildings and roads from high-resolution Maxar satellite imagery across the Arctic region. Shapefiles beginning with "HABITAT_AK" contain only the post-processed deep learning predictions. Shapefiles beginning with "HABITAT_OSM" contain the post-processed deep learning predictions supplemented by OpenStreetMap data. The HABITAT pipeline is based on a ResNet50-UNet++ semantic segmentation architecture trained on a training dataset comprised of building and road footprint polygons manually digitized from Maxar satellite imagery across the circumpolar Arctic (including Alaska, Russia, and Canada). The code is made available at https://github.com/PermafrostDiscoveryGateway/HABITAT. From imagery of 285 Alaskan communities acquired between 2018-2023, we detected approximately 250,000 buildings and storage tanks (comprising a 41.76 million square meter footprint) and 15 million meters of road. Building (including storage tanks) footprint polygons and road centerlines were strictly mapped within the boundaries of Alaskan communities (both incorporated places and census designated places). After the deep learning model detected building and road footprints, post-processing was performed to smooth out building footprints, extract centerlines from road footprints, and remove falsely-detected infrastructure. In particular, a buffer is created around developed land cover identified by the 2016 Alaska National Land Cover Database map, and model predictions that fall outside of the buffer are assumed to be confused with non-infrastructure land cover. Finally, we selected buildings and roads from the OpenStreetMap Alaska dataset (downloaded in June 2024 from https://download.geofabrik.de/) that do not intersect with any deep learning predictions to generate a merged OSM and HABITAT infrastructure dataset. This merged product comprises a total building footprint of 53 million square meters and a road network of 63,744 km across the state of Alaska.

We created HABITAT (High-resolution Arctic Built Infrastructure and Terrain Analysis Tool), a deep learning-based, high-performance computing-enabled mapping pipeline to automatically detect buildings and roads from high-resolution Maxar satellite imagery in Arctic communities. The code is made available at https://github.com/PermafrostDiscoveryGateway/HABITAT. The pipeline is based on a ResNet50-UNet++ semantic segmentation architecture trained on a training dataset comprised of building and road footprint polygons manually digitized from Maxar satellite imagery across the circumpolar Arctic (including Alaska, Russia, and Canada). From imagery of 285 Alaskan communities acquired between 2018-2023, we detected approximately 250,000 buildings and storage tanks (comprising a 41.76 million square meter footprint) and 15 million meters of road. Building (including storage tanks) footprint polygons and road centerlines were strictly mapped within the boundaries of Alaskan communities (both incorporated places and census designated places). After the deep learning model detected building and road footprints, post-processing was performed to smooth out building footprints, extract centerlines from road footprints, and remove falsely-detected infrastructure. In particular, a buffer is created around developed land cover identified by the 2016 Alaska National Land Cover Database map, and model predictions that fall outside of the buffer are assumed to be confused with non-infrastructure land cover.

The use of high-resolution remotely sensed imagery can be an effective way to obtain quantitative measurements of rock-avalanche volumes and geometries in remote glaciated areas, both of which are important for an improved understanding of rock-avalanche characteristics and processes. We utilized the availability of high-resolution (~0.5 m) WorldView satellite stereo imagery to derive digital elevation data in a 100 km2 area around the 28 June 2016 Lamplugh rock avalanche in Glacier Bay National Park and Preserve, Alaska. We used NASA Ames Stereo Pipeline, an open-source software package available from NASA, to produce one pre- and four post-event digital elevation models (DEMs) of the area surrounding the Lamplugh rock avalanche. This data release includes five raster elevation datasets (2-m resolution) in GeoTIFF format that have been orthrectified to the Universal Transverse Mercator (UTM) coordinate system (zone 7N). Elevations are measured in reference to the World Geodetic System 1984 (WGS84) ellipsoid. Because the study area is remote and difficult to access, ground control was not available to assess the absolute accuracy of DEMs. The DEMs have not been precisely co-registered. Data contained in this release include a pre-event DEM from 15 June 2016, and post-event DEMs from 16 July 2016, 27 August 2016, 27 September 2016, and 28 September 2016. The filenames for these DEMs are 20160615_LamplughDEM.tif, 20160716_LamplughDEM.tif, 20160827_LamplughDEM.tif, 20160927_LamplughDEM.tif, and 20160928_LamplughDEM.tif, respectively. We also provide a CSV file (Lamplugh_DEM_Image_Notes.csv) that contains the acquisition date, satellite platform, image identification number, resolution, off-nadir angle, and notes on image quality for each stereo pair used to generate DEMs.

The Earth Observation Satellite market is experiencing robust growth, driven by increasing demand for high-resolution imagery across diverse sectors. The market's expansion is fueled by several key factors. Firstly, advancements in sensor technology are enabling the capture of more detailed and accurate data, enhancing applications in infrastructure monitoring, environmental protection, and resource management. Secondly, the rising adoption of AI and machine learning for data analysis is unlocking new possibilities for extracting valuable insights from satellite imagery, improving decision-making across various industries. Thirdly, government initiatives promoting space exploration and environmental monitoring are fostering market growth by providing funding and encouraging technological advancements. Finally, the increasing affordability of satellite technology is making it accessible to a wider range of users, expanding market penetration. While challenges remain, such as the high initial investment costs associated with satellite development and launch, and potential regulatory hurdles, the overall market outlook is positive. The market segmentation reveals significant opportunities across different application areas. Infrastructure monitoring, encompassing urban planning, transportation, and construction, is a major driver of growth. Similarly, environmental monitoring, particularly for climate change studies, deforestation detection, and pollution control, is a rapidly expanding sector. The energy sector utilizes satellite imagery for exploration, resource management, and pipeline monitoring, while the natural resources sector leverages this technology for agriculture, mining, and forestry applications. Maritime surveillance, disaster management, and other niche applications further contribute to the market's diverse revenue streams. Geographically, North America and Europe currently hold significant market share, driven by robust technological advancements and high government investments. However, the Asia-Pacific region is projected to witness substantial growth in the coming years, fueled by increasing infrastructure development and rising government spending on space technology. This growth will likely be driven by countries such as China and India. Assuming a conservative CAGR of 8% and a 2025 market size of $15 billion, we can project a substantial increase over the forecast period.

We build a large-scale, comprehensive, and high-quality synthetic dataset for city-scale neural rendering researches. Leveraging the Unreal Engine 5 City Sample project, we developed a pipeline to easily collect aerial and street city views with ground-truth camera poses, as well as a series of additional data modalities. Flexible control on environmental factors like light, weather, human and car crowd is also available in our pipeline, supporting the need of various tasks covering city-scale neural rendering and beyond. The resulting pilot dataset, MatrixCity, contains 67k aerial images and 452k street images from two city maps of total size 28km^2.

A high-resolution, high vertical accuracy Digital Surface Model (DSM) of Mt. Etna was derived from Pleiades satellite data using the National Aeronautics and Space Administration (NASA) Ames Stereo Pipeline (ASP) tool set (https://ti.arc.nasa.gov/tech/asr/groups/intelligent-robotics/ngt/stereo/). The NASA Ames Stereo Pipeline (ASP) is a suite of free and open source automated geodesy and stereogrammetry tools designed for processing stereo imagery captured from satellites (around Earth and other planets), robotic rovers, aerial cameras, and historical imagery, with and without accurate camera pose information. The methodology used by the ASP software is similar with structure-from-motion (SfM) methodology using stereo triangulation combining spacecraft ephemeris/attitude information, sensor model, and disparity map information obtained by feature matching procedures. The model covers an area of about 400 km2 with a spatial resolution of 2 m and centers on the summit portion of th ...

The goal of the Alaska Advanced Very High Resolution Radiometer (AVHRR) project is to compile a time series data set of calibrated, georegistered daily observations and twice-monthly maximum normalized difference vegetation index (NDVI) composites for Alaska's annual growing season (April-October). This data set has applications for environmental monitoring and for assessing impacts of global climate change. An Alaska AVHRR data set is comprised of twice-monthly maximum NDVI composites of daily satellite observations. The NDVI composites contain 10 bands of information, including AVHRR channels 1-5, maximum NDVI, satellite zenith, solar zenith, and relative azimuth. The daily observations, bands 1-9, have been calibrated to reflectance, scaled to byte data, and geometrically registered to the Albers Equal-Area Conic map projection. The 10th band is a pointer to identify the date and scene ID of the source daily observation (scene) for each pixel.

The compositing process required each daily overpass to be registered to a common map projection to ensure that from day to day each 1-km pixel represented the exact same ground location. The Albers Equal-Area Conic map projection provides for equal area representation, which enables easy measurement of area throughout the data. Each daily observation for the growing season was registered to a base image using image-to-image correlation.

The NDVI data are calculated from the calibrated, geometrically registered daily observations. The NDVI value is the difference between near-infrared (AVHRR Channel 2) and visible (AVHRR Channel 1) reflectance values divided by total measured reflectance. A maximum NDVI compositing process was used on the daily observations. The NDVI is examined pixel by pixel for each observation during the compositing period to determine and retain the maximum value. Often when displaying data covering large areas, such as AVHRR data, it is beneficial to include an overlay of either familiar linework for reflectance or polygon data sets to derive statistical summaries of regions. All of the linework images represent lines in raster format as 1-km cells and the strata are represented as polygons registered to the AVHRR data. The linework and polygon data sets include international boundaries, Alaskan roads with the Trans-Alaska Pipeline, and a raster polygon mask of the State.

FSC-180k

We introduce our hybrid semantic change detection dataset, named FSC-180k. It consists of approximately 60,000 real aerial images sourced from the FLAIR dataset, along with 180,000 artificially modified images. These images were generated using our HySCDG pipeline applied (three times) to each real image. In total, the dataset provides 180,000 image pairs. Each pair is accompanied by a binary change map and semantic segmentation maps for both images (land use… See the full description on the dataset page: https://huggingface.co/datasets/Yanis236/fsc-180k.

Photovoltaic (PV) energy generation plays a crucial role in the energy transition. Small-scale PV installations are deployed at an unprecedented pace, and their integration into the grid can be challenging since stakeholders often lack quality data about these installations. Overhead imagery is increasingly used to improve the knowledge of distributed PV installations with machine learning models capable of automatically mapping these installations. However, these models cannot be easily transferred from one region or data source to another due to differences in image acquisition. To address this issue known as domain shift and foster the development of PV array mapping pipelines, we propose a dataset containing aerial images, annotations, and segmentation masks. We provide installation metadata for more than 28,000 installations. We provide ground truth segmentation masks for 13,000 installations, including 7,000 with annotations for two different image providers. Finally, we provide ground truth annotations and associated installation metadata for more than 8,000 installations. Dataset applications include end-to-end PV registry construction, robust PV installations mapping, and analysis of crowdsourced datasets.

This dataset contains the complete records associated with the article "A crowdsourced dataset of aerial images of solar panels, their segmentation masks, and characteristics", currently under review. These complete records consist of RGB overhead imagery, segmentation masks, and characteristics of PV installations. The data records are organized as follows:

• bdappv/ Root data folder
  • google / ign: One folder for each campaign
    • img/: Folder containing all the images presented to the users. This folder contains 28807 images for Google and 17325 images for IGN.
    • mask/: Folder containing all segmentations masks generated from the polygon annotations of the users. This folder contains 13303 masks for Google and 7686 masks for IGN.
• metadata.csv The .csv file with the characteristics of the installations.

Logged forests cover four million square kilometers of the tropics, capturing carbon more rapidly than temperate forests and harboring rich biodiversity. Restoring these forests is essential to help avoid the worst impacts of climate change. Yet monitoring tropical forest recovery is challenging. We track the abundance of early-successional species in a forest restoration concession in Indonesia. If the species are carefully chosen, they can be used as an indicator of restoration progress. We present SLIC-UAV, a new pipeline for processing Unoccupied Aerial Vehicle (UAV) imagery using simple linear iterative clustering (SLIC)to map early-successional species in tropical forests. The pipeline comprises: (a) a field verified approach for manually labeling species; (b) automatic segmentation of imagery into “superpixels” and (c) machine learning classification of species based on both spectral and textural features. Creating superpixels massively reduces the dataset's dimensionality and enables the use of textural features, which improve classification accuracy. In addition, this approach is flexible with regards to the spatial distribution of training data. This allowed us to be flexible in the field and collect high-quality training data with the help of local experts. The accuracy ranged from 74.3% for a four-species classification task to 91.7% when focusing only on the key early-succesional species. We then extended these models across 100 hectares of forest, mapping species dominance and forest condition across the entire restoration project.

Evaluation of the the Segment Anything Model (SAM) for penguin colony segmentation using mean intersection over union (mIoU), difference in perimeter to area ratio (PAR), area error, and accuracy (i.e. panels a-c in Figs 3 and 4 vs. ground truth). 95% confidence intervals are shown. An up (down) arrow indicates a measure where a larger (smaller) number is preferred.

Geospecific View Generation - Geometry-Context Aware High-resolution Ground View Inference from Satellite Views

🌐 Homepage | 📖 arXiv

  Introduction

Predicting realistic ground views from satellite imagery in urban scenes is a challenging task due to the significant view gaps between satellite and ground-view images. We propose a novel pipeline to tackle this challenge, by generating geospecifc views that maximally respect the weak geometry and texture from multi-view… See the full description on the dataset page: https://huggingface.co/datasets/GDAlab/GeoContext-v1.

Class I4 easements relate to easements in the vicinity of an overhead or underground power line. These are two categories of servitudes established by the Act of 15 June 1906 on energy distributions. the easements provided for in paragraphs 1, 2, 3 and 4 of section 12 concerning all electrical energy distributions: — anchorage easement to permanently establish supports and anchorages for overhead electric conductors, either outside the walls or facades overlooking the highway, or on the roofs and terraces of buildings, — overhanging easement allowing electric conductors to pass over private property, — easement of passage or support enabling underground pipes to be permanently established, or support for aerial operators, on unbuilt private land, which is not closed by walls or other equivalent fences, — tree pruning and cutting easement to cut trees and branches of trees which, in the vicinity of overhead electric conductors, hinder their installation or could, by their movement or fall, cause short circuits or damage to the structures. These are easements that do not result in any dispossession of the owner who retains the right to demolish, repair, raise, close or build, subject to notifying the concessionaire one month before commencing the work. perimeters established pursuant to Article 12 bis on either side of an overhead power line with a voltage of 130 kilovolts or more and within which: — the following are prohibited: • residential buildings, • Traveller reception areas, • certain categories of establishments receiving from the public: reception facilities for the elderly and disabled persons, hotels and accommodation facilities, educational establishments, holiday camps, health facilities, penitentiary establishments, outdoor establishments. — may be prohibited or subject to requirements: • other categories of establishments receiving from the public, • installations classified for the protection of the environment subject to authorisation and manufacturing, using or storing oxidising, explosive, flammable or combustible substances, without, however, hindering the adaptation, rehabilitation or extension of existing ones, provided, however, that the capacity of inhabitants within the area of servitudes may not be increased. This resource describes the surface bases of Class I4 easements combined with their generators, i.e. all electrical power distribution facilities, including: — overhead electric conductors, — underground electricity transmission pipelines, — aerial conductors, — structures, such as processing stations, etc.

Source: —NR— Vintage: —NR— Dissemination: Restricted

Evaluation of final predicted penguin colony areas at Devil Island using mean intersection over union (mIoU), difference in perimeter to area ratio (PAR), area error, and accuracy (i.e. Fig 5 vs. ground truth). 95% confidence intervals are shown. We also show the evaluation of a fully manual approach. An up (down) arrow indicates a measure where a larger (smaller) number is preferred.

This data set contains shapefiles of termini traces from 294 Greenland glaciers, derived using a deep learning algorithm (AutoTerm) applied to satellite imagery. The model functions as a pipeline, imputing publicly availably satellite imagery from Google Earth Engine (GEE) and outputting shapefiles of glacial termini positions for each image. Also available are supplementary data, including temporal coverage of termini traces, time series data of termini variations, and updated land, ocean, and ice masks derived from the Greenland Ice Sheet Mapping Project (GrIMP) ice masks.

=========================== Authors Etienne BERTHIER LEGOS, Université de Toulouse, CNES, CNRS, IRD, UPS, 31400 Toulouse, France, =========================== 1. Collection This collection contains three digital elevation models (DEMs) of "Little Kluane" Glacier (Yukon Territory, Canada) with an horizontal grid spacing of 30 m * SPOT5, 13 September 2007 * ASTER, 30 September 2017 * SPOT7, 1 October 2018 The original SPOT5 DEM was obtained from the SPIRIT project (Korona et al., 2009) ASTER and SPOT7 DEMs have been derived using the Ames Stereo Pipeline (Beyer et al., 2018) from stereo images using the set of correlation of parameters in Deschamps-Berger et al. (2020) All DEMs have been coregistered and bias-corrected to the Copernicus 30 m global DEM following the methods of Berthier & Brun (2019) =========================== 2. Dataset Acknowledgement * SPOT5 data were obtained thanks to funding from CNES during the fourth international polar year. * SPOT7 data were obtained thanks to DINAMIS Project, for “Dispositif Institutionnel National d’Approvisionnement Mutualisé en Imagerie Satellitaire”, a French platform that acquires and distributes very high resolution Earth satellite imagery for French and foreign institutional users under specific subscription conditions. * ASTER data were provided by NASA (U.S.) and METI (Japan). =========================== 3. Dataset Attribution This dataset is licensed under a Creative Commons CC BY-NC 4.0 International License (Attribution-NonCommercial). =========================== 4. Related publication This dataset has been generated for and used in a publication to be submitted to the Journal of Glaciology : Morin, A., Flowers, G. E., Nolan, A., Brinkerhoff, D. J., and Berthier, E.: Exploiting high-slip ?ow regimes to improve bed inference, submitted. =========================== 5. Collection Location Yukon Territory of Canada Bounding box: WGS 84 / UTM zone 7N =========================== References Berthier, E. and Brun, F.: Karakoram geodetic glacier mass balances between 2008 and 2016: persistence of the anomaly and influence of a large rock avalanche on Siachen Glacier, J Glaciol, 65, 494–507, https://doi.org/10.1017/jog.2019.32, 2019. Beyer, R. A., Alexandrov, O., and McMichael, S.: The Ames Stereo Pipeline: NASA’s Open Source Software for Deriving and Processing Terrain Data, Earth and Space Science, 5, 537–548, https://doi.org/10.1029/2018EA000409, 2018. Deschamps-Berger, C., Gascoin, S., Berthier, E., Deems, J., Gutmann, E., Dehecq, A., Shean, D., and Dumont, M.: Snow depth mapping from stereo satellite imagery in mountainous terrain: evaluation using airborne laser-scanning data, The Cryosphere, 14, 2925–2940, https://doi.org/10.5194/tc-14-2925-2020, 2020. Korona, J., Berthier, E., Bernard, M., Remy, F., and Thouvenot, E.: SPIRIT. SPOT 5 stereoscopic survey of Polar Ice: Reference Images and Topographies during the fourth International Polar Year (2007-2009), ISPRS J. Photogramm., 64, 204–212, https://doi.org/10.1016/j.isprsjprs.2008.10.005, 2009.

1- Dataset Summary
Here we present a dataset of DEMs (Digital Elevation Models), orthomosaics, and lava area outlines for the August 2022 eruption at Fagradalsfjall, SW Iceland. The dataset consists of: (1) five aerial surveys collected over the course of the August 2022 Fagradalsfjall eruption, (2) one survey carried out on 14 August 2022 using Pléiades satellite stereo images, and (3) a larger aerial survey, covering the 2021 and 2022 eruption sites in late September 2022 after the volcanic activity concluded.

2- Background

The volcano at Fagradalsfjall, SW-Iceland, began erupting on 3 August 2022 at 13:20 following 10 months of quiescence. As part of the response plan, a series of photogrammetric surveys were conducted in rapid, operational mode throughout the duration of the eruption. Subsequent production of data products for natural hazards monitoring (lava maps, lava volumes, effusion rates) were calculated within hours and reported to the Icelandic Civil Defense, following a similar approach that described in Pedersen et al., 2022a and in Gouhier et al., 2022. At the start of the 2022 eruption, GCPs had not yet been placed around the new fissure, but reference data (orthomosaics and DEMs) which had been georeferenced using targets measured with differential GNSS existed of the eruption site from September 2021 from Pedersen et al. (2022b) were available to use as a reference in the new workflow instead of GCPs. Due to the urgent need from authorities for information about the new eruption, a processing method that avoids the time-consuming task of manual GCP selection using a reference image for georeferencing was preferable in this instance. Besides the acquisition of aerial photographs, the CIEST2 initiative was also re-activated to collect Pléiades stereo images in emergency mode (Gouhier et al., 2022).

3 – Overview of data collection

Table 1 contains the overview of the surveys collected and presented in this repository.

Table 1. Summary of surveys included in this dataset, by survey date.

* TF-203: Savannah S aircraft

** TF-BMW: Vulcanair P68 Observer 2 aircraft, operated by Garðaflug ehf.

4- Methods

4.1 Processing of the aerial photographs from 3-16 Aug 2022
Throughout the eruption, aerial surveys were conducted using a Hasselblad A6D 100 MP camera with 35 mm focal lens, from a height of 750 – 2,100 m above ground over the active lava field from an ultralight aircraft with a window in the bottom to allow for vertical photos to be taken (see supplement of Pedersen et al., 2022a for details and images of the setup). The camera was manually triggered to give ~70% overlap, and approximate flight lines were prepared beforehand for use with a handheld GPS during the flight to give ~30 % side overlap.

An automated processing pipeline was created in python, which leverages tools from the Ames Stereo Pipeline (ASP, Shean et al., 2016) and Agisoft Metashape stand-alone Python API (v. 1.8.4). The processing and georeferencing of the aerial data were done in three steps, with all steps being automated except for the digitization of lava outlines. First, using a very high-resolution reference orthomosaic and DEM created in September 2021 and georeferenced with ground control points (Pedersen et al., 2022b), interest points (IPs) in each image were matched with the reference dataset, using the ASP routine ipfind. This created GCPs for each image over stable terrain. Second, hillshades were created from both the reference DEM and the source dataset DEM and matches in IPs were found in both, creating a second round of ground control points to refine the georeferencing of the entire block. Finally, the alignment of the source DEM was refined using the dem_align (demcoreg) protocol from Shean et al. (2016) by applying a bulk linear shift in X, Y and Z which minimizes the vertical difference in stable terrain between the source and reference DEM.

4.2 Processing of the Pléiades stereo images
The Pléiades stereo images were processed using the Ames Stereo Pipeline, using the general workflow of mapproject and parallel_stereo (e.g., Deschamps-Berger et al., 2020). The parallel_stereo routine used default arguments, plus the following arguments:

--stereo-algorithm asp_mgm -t rpcmaprpc --corr-seed-mode 3 --corr-max-levels 2 --cost-mode 3 --subpixel-mode 9 --corr-kernel 7 7 --subpixel-kernel 15 15

We used the DEM from 4 Aug 2022 as the reference for mapproject and for the final DEM co-registration applied to the produced Pléiades DEM.

4.3 Processing of the 26 September 2022 dataset
The survey from 26 September 2022 was collected and processed using direct georeferencing from an on-board GPS antenna. The final alignment of the block was refined using the dem_align (demcoreg) protocol from Shean et al. (2016) by applying a bulk linear shift in X, Y and Z which minimizes the vertical difference in stable terrain between the source and reference DEM. Because this survey covered a much larger area, the reference DEM for the final coregistration was the ÍslandsDEM v.1.0 (Landmælingar Íslands, 2022).

4.4. Maps of the lava outlines, lava thickness, lava volume, Time Average Effusion Rate (TADR)
For each survey, a differential DEM (dDEM) showing elevation changes since the 2021 eruption was created by subtracting the reference DEM (ÍslandsDEM v.1.0, which includes the post-eruption DEM from Pedersen et al., 2022a) from the source DEM. Lava outlines, lava thickness lava volume, TADR and uncertainties were calculated using the methods described in Pedersen et al., 2022a. Table 2 summarizes calculations from this dataset.

Table 2. Summary of survey results calculated from August 2022 Fagradalsfjall eruption DEMs and orthomosaics.

    <p><strong>Difference</strong></p>
    </th>
    <th>
    <p><strong>Lava </strong></p>

    <p><strong>Area<sup>*</sup> End </strong></p>

    <p><strong>(km<sup>2</sup>)</strong></p>
    </th>
    <th>
    <p><strong>dh<sup>**</sup> </strong></p>

    <p><strong>(m)</strong></p>
    </th>
    <th>
    <p><strong>Volume<sup>+</sup></strong></p>

    <p><strong>End<br>
    (1e+6 m<sup>3</sup>)</strong></p>
    </th>
    <th>
    <p><strong>TADR<sup>++</sup><br>
    (m<sup>3</sup>/s)</strong></p>
    </th>
  </tr>
</thead>
<tbody>
  <tr>
    <td>
    <p>20220803<br>
    13:20</p>
    </td>
    <td>
    <p>20220803<br>
    17:05</p>
    </td>
    <td>
    <p>0d 03h 45m</p>
    </td>
    <td>
    <p>0.07</p>
    </td>
    <td>
    <p>5.88</p>
    </td>
    <td>
    <p>0.43</p>

    <p>± 0.03</p>
    </td>
    <td>
    <p>32.1</p>

    <p>± 1.5</p>
    </td>
  </tr>
  <tr>
    <td>
    <p>20220803<br>
    17:05</p>
    </td>
    <td>
    <p>20220804<br>
    11:00</p>
    </td>
    <td>
    <p>0d 21h 40m</p>
    </td>
    <td>
    <p>0.14</p>
    </td>
    <td>
    <p>11.13 </p>
    </td>
    <td>
    <p>1.57</p>

    <p>± 0.05</p>
    </td>
    <td>
    <p>17.7</p>

    <p>± 0.8</p>
    </td>
  </tr>
  <tr>
    <td>
    <p>20220804<br>
    11:00</p>
    </td>
    <td>
    <p>20220813<br>
    09:00</p>
    </td>
    <td>
    <p>8d 22h 00m</p>
    </td>
    <td>
    <p>1.27<sup>#</sup></p>
    </td>
    <td>
    <p>6.90</p>
    </td>
    <td>
    <p>10.33</p>

    <p>± 0.6</p>
    </td>
    <td>
    <p>11.4</p>

    <p>± 0.7</p>
    </td>
  </tr>
  <tr>
    <td>
    <p>20220813<br>
    13:08</p>
    </td>
    <td>
    <p>20220814<br>
    13:00</p>
    </td>
    <td>
    <p>0d 23h

Doodleverse/Segmentation Zoo/Seg2Map Res-UNet models for segmentation of buildings of RGB 1024x1024 high-res. images

Models have been created using Segmentation Gym* using the following dataset**: https://github.com/FrontierDevelopmentLab/multi3net

These Residual-UNet model data are based on 1m spatial footprint images and associated labels of buildings in Houston. Imagery made available through DigitalGlobe***

Image size used by model: 1024 x 1024 x 3 pixels

classes:
other
building

File descriptions

For each model, there are 5 files with the same root name:

1. '.json' config file: this is the file that was used by Segmentation Gym* to create the weights file. It contains instructions for how to make the model and the data it used, as well as instructions for how to use the model for prediction. It is a handy wee thing and mastering it means mastering the entire Doodleverse.

2. '.h5' weights file: this is the file that was created by the Segmentation Gym* function `train_model.py`. It contains the trained model's parameter weights. It can called by the Segmentation Gym* function `seg_images_in_folder.py`. Models may be ensembled.

3. '_modelcard.json' model card file: this is a json file containing fields that collectively describe the model origins, training choices, and dataset that the model is based upon. There is some redundancy between this file and the `config` file (described above) that contains the instructions for the model training and implementation. The model card file is not used by the program but is important metadata so it is important to keep with the other files that collectively make the model and is such is considered part of the model

4. '_model_history.npz' model training history file: this numpy archive file contains numpy arrays describing the training and validation losses and metrics. It is created by the Segmentation Gym function `train_model.py`

5. '.png' model training loss and mean IoU plot: this png file contains plots of training and validation losses and mean IoU scores during model training. A subset of data inside the .npz file. It is created by the Segmentation Gym function `train_model.py`

Additionally, BEST_MODEL.txt contains the name of the model with the best validation loss and mean IoU

References
*Segmentation Gym: Buscombe, D., & Goldstein, E. B. (2022). A reproducible and reusable pipeline for segmentation of geoscientific imagery. Earth and Space Science, 9, e2022EA002332. https://doi.org/10.1029/2022EA002332 See: https://github.com/Doodleverse/segmentation_gym

**Rudner, T. G. J.; Rußwurm, M.; Fil, J.; Pelich, R.; Bischke, B.; Kopačková, V.; Biliński, P. Segmenting Flooded Buildings via Fusion of Multiresolution, Multisensor, and Multitemporal Satellite Imagery. In AAAI 2019. https://arxiv.org/pdf/1812.01756.pdf

***DigitalGlobe. 2018. DigitalGlobe Open Data Program. https://www.digitalglobe.com/opendata. Online; accessed 2018-09-01.