Find research datasets worth reusing

Underwater images collected by an Autonomous Surface Vehicle in Trou-Deau, Réunion - 2024-05-17

<i>This dataset was collected by an Autonomous Surface Vehicle in Trou-Deau, Réunion - 2024-05-17.</i> <br> <br><br>Underwater or aerial images collected by scientists or citizens can have a wide variety of use for science, management, or conservation. These images can be annotated and shared to train IA models which can in turn predict the objects on the images. We provide a set of tools (hardware and software) to collect marine data, predict species or habitat, and provide maps.<br><br> This dataset is part of larger collection referencing numerous underwater and aerial images <a href="https://doi.org/10.5281/zenodo.11125847" target="_blank">Seatizen Altas</a>. Methods, tools and scientific objectives are also described in a dedicated data paper.<br> <h2>Image acquisition</h2> This session has 29.02 GB of MP4 files, which were trimmed into 9488 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 99.28% of these extracted images are useful and 0.72% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. <br> Multilabel predictions have been made on useful frames using <a href="https://huggingface.co/lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze" target="_blank">DinoVd'eau</a> model. <br> <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. <br> Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 56.41 %, Q2: 43.46 %, Q5: 0.13 % <br> <h2> Generic folder structure </h2> YYYYMMDD_COUNTRYCODE-optionalplace_device_session-number <br> ├── DCIM : folder to store videos and photos depending on the media collected. <br> ├── GPS : folder to store any positioning related file. If any kind of correction is possible on files (e.g. Post-Processed Kinematic thanks to rinex data) then the distinction between device data and base data is made. If, on the other hand, only device position data are present and the files cannot be corrected by post-processing techniques (e.g. gpx files), then the distinction between base and device is not made and the files are placed directly at the root of the GPS folder. <br> │ ├── BASE : files coming from rtk station or any static positioning instrument. <br> │ └── DEVICE : files coming from the device. <br> ├── METADATA : folder with general information files about the session. <br> ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. <br> │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. <br> │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. <br> │ ├── IA : destination folder for image recognition predictions. <br> │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. <br> └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <br> <h2> Software </h2> All the raw data was processed using our <a href="https://github.com/SeatizenDOI/plancha-workflow/releases/tag/v1.0.3" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. <br>You can find all the necessary scripts to download this data in this <a href="https://github.com/SeatizenDOI/zenodo-tools" target="_blank">repository</a>. <br>Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>! <br>

Underwater images collected by an Autonomous Surface Vehicle in St-Leu, Réunion - 2023-11-10

<i>This dataset was collected by an Autonomous Surface Vehicle in St-Leu, Réunion - 2023-11-10.</i> <br> <br><br>Underwater or aerial images collected by scientists or citizens can have a wide variety of use for science, management, or conservation. These images can be annotated and shared to train IA models which can in turn predict the objects on the images. We provide a set of tools (hardware and software) to collect marine data, predict species or habitat, and provide maps.<br><br> This dataset is part of larger collection referencing numerous underwater and aerial images <a href="https://doi.org/10.5281/zenodo.11125847" target="_blank">Seatizen Altas</a>. Methods, tools and scientific objectives are also described in a dedicated data paper.<br> <h2>Image acquisition</h2> This session has 24.5 GB of MP4 files, which were trimmed into 8346 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 100.0% of these extracted images are useful and 0.0% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. <br> Multilabel predictions have been made on useful frames using <a href="https://huggingface.co/lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze" target="_blank">DinoVd'eau</a> model. <br> <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. <br> Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 97.18 %, Q2: 2.67 %, Q5: 0.14 % <br> <h2> Bathymetry </h2> The data are collected using a single-beam echosounder <a href="https://www.echologger.com/products/single-frequency-echosounder-deep" target="_blank">ETC 400</a>. <br> We only keep the values which have a GPS correction in Q1.<br> We keep the points that are the waypoints.<br> We keep the raw data where depth was estimated between 0.2 m and 50.0 m deep. <br> The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available.<br> At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. <br> The size of the grid cells is 0.19 m. <br> The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <br> <h2> Generic folder structure </h2> YYYYMMDD_COUNTRYCODE-optionalplace_device_session-number <br> ├── DCIM : folder to store videos and photos depending on the media collected. <br> ├── GPS : folder to store any positioning related file. If any kind of correction is possible on files (e.g. Post-Processed Kinematic thanks to rinex data) then the distinction between device data and base data is made. If, on the other hand, only device position data are present and the files cannot be corrected by post-processing techniques (e.g. gpx files), then the distinction between base and device is not made and the files are placed directly at the root of the GPS folder. <br> │ ├── BASE : files coming from rtk station or any static positioning instrument. <br> │ └── DEVICE : files coming from the device. <br> ├── METADATA : folder with general information files about the session. <br> ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. <br> │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. <br> │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. <br> │ ├── IA : destination folder for image recognition predictions. <br> │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. <br> └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <br> <h2> Software </h2> All the raw data was processed using our <a href="https://github.com/SeatizenDOI/plancha-workflow/releases/tag/v1.0.3" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. <br>You can find all the necessary scripts to download this data in this <a href="https://github.com/SeatizenDOI/zenodo-tools" target="_blank">repository</a>. <br>Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>! <br>

Underwater images collected by an Autonomous Surface Vehicle in Hermitage, Réunion - 2024-06-28

<i>This dataset was collected by an Autonomous Surface Vehicle in Hermitage, Réunion - 2024-06-28.</i> <br> <br><br>Underwater or aerial images collected by scientists or citizens can have a wide variety of use for science, management, or conservation. These images can be annotated and shared to train IA models which can in turn predict the objects on the images. We provide a set of tools (hardware and software) to collect marine data, predict species or habitat, and provide maps.<br><br> This dataset is part of larger collection referencing numerous underwater and aerial images <a href="https://doi.org/10.5281/zenodo.11125847" target="_blank">Seatizen Altas</a>. Methods, tools and scientific objectives are also described in a dedicated data paper.<br> <h2>Image acquisition</h2> This session has 27.79 GB of MP4 files, which were trimmed into 9234 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 99.57% of these extracted images are useful and 0.43% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. <br> Multilabel predictions have been made on useful frames using <a href="https://huggingface.co/lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze" target="_blank">DinoVd'eau</a> model. <br> <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. <br> Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 91.11 %, Q2: 6.53 %, Q5: 2.36 % <br> <h2> Bathymetry </h2> The data are collected using a single-beam echosounder <a href="https://www.echologger.com/products/single-frequency-echosounder-deep" target="_blank">ETC 400</a>. <br> We only keep the values which have a GPS correction in Q1.<br> We keep the points that are the waypoints.<br> We keep the raw data where depth was estimated between 0.2 m and 50.0 m deep. <br> The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available.<br> At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. <br> The size of the grid cells is 0.217 m. <br> The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <br> <h2> Generic folder structure </h2> YYYYMMDD_COUNTRYCODE-optionalplace_device_session-number <br> ├── DCIM : folder to store videos and photos depending on the media collected. <br> ├── GPS : folder to store any positioning related file. If any kind of correction is possible on files (e.g. Post-Processed Kinematic thanks to rinex data) then the distinction between device data and base data is made. If, on the other hand, only device position data are present and the files cannot be corrected by post-processing techniques (e.g. gpx files), then the distinction between base and device is not made and the files are placed directly at the root of the GPS folder. <br> │ ├── BASE : files coming from rtk station or any static positioning instrument. <br> │ └── DEVICE : files coming from the device. <br> ├── METADATA : folder with general information files about the session. <br> ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. <br> │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. <br> │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. <br> │ ├── IA : destination folder for image recognition predictions. <br> │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. <br> └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <br> <h2> Software </h2> All the raw data was processed using our <a href="https://github.com/SeatizenDOI/plancha-workflow/releases/tag/v1.0.3" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. <br>You can find all the necessary scripts to download this data in this <a href="https://github.com/SeatizenDOI/zenodo-tools" target="_blank">repository</a>. <br>Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>! <br>

Underwater images collected by an Autonomous Surface Vehicle in Hermitage, Réunion - 2024-06-28

<i>This dataset was collected by an Autonomous Surface Vehicle in Hermitage, Réunion - 2024-06-28.</i> <br> <br><br>Underwater or aerial images collected by scientists or citizens can have a wide variety of use for science, management, or conservation. These images can be annotated and shared to train IA models which can in turn predict the objects on the images. We provide a set of tools (hardware and software) to collect marine data, predict species or habitat, and provide maps.<br><br> This dataset is part of larger collection referencing numerous underwater and aerial images <a href="https://doi.org/10.5281/zenodo.11125847" target="_blank">Seatizen Altas</a>. Methods, tools and scientific objectives are also described in a dedicated data paper.<br> <h2>Image acquisition</h2> This session has 29.62 GB of MP4 files, which were trimmed into 9962 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 99.97% of these extracted images are useful and 0.03% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. <br> Multilabel predictions have been made on useful frames using <a href="https://huggingface.co/lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze" target="_blank">DinoVd'eau</a> model. <br> <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. <br> Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 87.68 %, Q2: 5.45 %, Q5: 6.87 % <br> <h2> Bathymetry </h2> The data are collected using a single-beam echosounder <a href="https://www.echologger.com/products/single-frequency-echosounder-deep" target="_blank">ETC 400</a>. <br> We only keep the values which have a GPS correction in Q1.<br> We keep the points that are the waypoints.<br> We keep the raw data where depth was estimated between 0.2 m and 50.0 m deep. <br> The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available.<br> At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. <br> The size of the grid cells is 0.141 m. <br> The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <br> <h2> Generic folder structure </h2> YYYYMMDD_COUNTRYCODE-optionalplace_device_session-number <br> ├── DCIM : folder to store videos and photos depending on the media collected. <br> ├── GPS : folder to store any positioning related file. If any kind of correction is possible on files (e.g. Post-Processed Kinematic thanks to rinex data) then the distinction between device data and base data is made. If, on the other hand, only device position data are present and the files cannot be corrected by post-processing techniques (e.g. gpx files), then the distinction between base and device is not made and the files are placed directly at the root of the GPS folder. <br> │ ├── BASE : files coming from rtk station or any static positioning instrument. <br> │ └── DEVICE : files coming from the device. <br> ├── METADATA : folder with general information files about the session. <br> ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. <br> │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. <br> │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. <br> │ ├── IA : destination folder for image recognition predictions. <br> │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. <br> └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <br> <h2> Software </h2> All the raw data was processed using our <a href="https://github.com/SeatizenDOI/plancha-workflow/releases/tag/v1.0.3" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. <br>You can find all the necessary scripts to download this data in this <a href="https://github.com/SeatizenDOI/zenodo-tools" target="_blank">repository</a>. <br>Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>! <br>

Underwater images collected by an Autonomous Surface Vehicle in Tessier, Réunion - 2024-05-15

<i>This dataset was collected by an Autonomous Surface Vehicle in Tessier, Réunion - 2024-05-15.</i> <br> <br><br>Underwater or aerial images collected by scientists or citizens can have a wide variety of use for science, management, or conservation. These images can be annotated and shared to train IA models which can in turn predict the objects on the images. We provide a set of tools (hardware and software) to collect marine data, predict species or habitat, and provide maps.<br><br> This dataset is part of larger collection referencing numerous underwater and aerial images <a href="https://doi.org/10.5281/zenodo.11125847" target="_blank">Seatizen Altas</a>. Methods, tools and scientific objectives are also described in a dedicated data paper.<br> <h2>Image acquisition</h2> This session has 19.42 GB of MP4 files, which were trimmed into 6576 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 99.7% of these extracted images are useful and 0.3% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. <br> Multilabel predictions have been made on useful frames using <a href="https://huggingface.co/lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze" target="_blank">DinoVd'eau</a> model. <br> <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. <br> Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 23.49 %, Q2: 75.4 %, Q5: 1.1 % <br> <h2> Generic folder structure </h2> YYYYMMDD_COUNTRYCODE-optionalplace_device_session-number <br> ├── DCIM : folder to store videos and photos depending on the media collected. <br> ├── GPS : folder to store any positioning related file. If any kind of correction is possible on files (e.g. Post-Processed Kinematic thanks to rinex data) then the distinction between device data and base data is made. If, on the other hand, only device position data are present and the files cannot be corrected by post-processing techniques (e.g. gpx files), then the distinction between base and device is not made and the files are placed directly at the root of the GPS folder. <br> │ ├── BASE : files coming from rtk station or any static positioning instrument. <br> │ └── DEVICE : files coming from the device. <br> ├── METADATA : folder with general information files about the session. <br> ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. <br> │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. <br> │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. <br> │ ├── IA : destination folder for image recognition predictions. <br> │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. <br> └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <br> <h2> Software </h2> All the raw data was processed using our <a href="https://github.com/SeatizenDOI/plancha-workflow/releases/tag/v1.0.3" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. <br>You can find all the necessary scripts to download this data in this <a href="https://github.com/SeatizenDOI/zenodo-tools" target="_blank">repository</a>. <br>Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>! <br>

Underwater images collected by an Autonomous Surface Vehicle in Pointe-Des-Aigrettes, Réunion - 2024-07-08

<i>This dataset was collected by an Autonomous Surface Vehicle in Pointe-Des-Aigrettes, Réunion - 2024-07-08.</i> <br> <br><br>Underwater or aerial images collected by scientists or citizens can have a wide variety of use for science, management, or conservation. These images can be annotated and shared to train IA models which can in turn predict the objects on the images. We provide a set of tools (hardware and software) to collect marine data, predict species or habitat, and provide maps.<br><br> This dataset is part of larger collection referencing numerous underwater and aerial images <a href="https://doi.org/10.5281/zenodo.11125847" target="_blank">Seatizen Altas</a>. Methods, tools and scientific objectives are also described in a dedicated data paper.<br> <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. <br> Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 91.26 %, Q2: 2.39 %, Q5: 6.35 % <br> <h2> Bathymetry </h2> The data are collected using a single-beam echosounder <a href="https://www.echologger.com/products/single-frequency-echosounder-deep" target="_blank">ETC 400</a>. <br> We only keep the values which have a GPS correction in Q1.<br> We keep the points that are the waypoints.<br> We keep the raw data where depth was estimated between 3.0 m and 50.0 m deep. <br> The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available.<br> At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. <br> The size of the grid cells is 0.733 m. <br> The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <br> <h2> Generic folder structure </h2> YYYYMMDD_COUNTRYCODE-optionalplace_device_session-number <br> ├── DCIM : folder to store videos and photos depending on the media collected. <br> ├── GPS : folder to store any positioning related file. If any kind of correction is possible on files (e.g. Post-Processed Kinematic thanks to rinex data) then the distinction between device data and base data is made. If, on the other hand, only device position data are present and the files cannot be corrected by post-processing techniques (e.g. gpx files), then the distinction between base and device is not made and the files are placed directly at the root of the GPS folder. <br> │ ├── BASE : files coming from rtk station or any static positioning instrument. <br> │ └── DEVICE : files coming from the device. <br> ├── METADATA : folder with general information files about the session. <br> ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. <br> │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. <br> │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. <br> │ ├── IA : destination folder for image recognition predictions. <br> │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. <br> └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <br> <h2> Software </h2> All the raw data was processed using our <a href="https://github.com/SeatizenDOI/plancha-workflow/releases/tag/v1.0.3" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. <br>You can find all the necessary scripts to download this data in this <a href="https://github.com/SeatizenDOI/zenodo-tools" target="_blank">repository</a>. <br>Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>! <br>

Underwater images collected by an Autonomous Surface Vehicle in Hermitage, Réunion - 2023-11-20

<i>This dataset was collected by an Autonomous Surface Vehicle in Hermitage, Réunion - 2023-11-20.</i> <br> <br><br>Underwater or aerial images collected by scientists or citizens can have a wide variety of use for science, management, or conservation. These images can be annotated and shared to train IA models which can in turn predict the objects on the images. We provide a set of tools (hardware and software) to collect marine data, predict species or habitat, and provide maps.<br><br> This dataset is part of larger collection referencing numerous underwater and aerial images <a href="https://doi.org/10.5281/zenodo.11125847" target="_blank">Seatizen Altas</a>. Methods, tools and scientific objectives are also described in a dedicated data paper.<br> <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. <br> Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 83.45 %, Q2: 16.28 %, Q5: 0.27 % <br> <h2> Bathymetry </h2> The data are collected using a single-beam echosounder <a href="https://www.echologger.com/products/single-frequency-echosounder-deep" target="_blank">ETC 400</a>. <br> We only keep the values which have a GPS correction in Q1.<br> We keep the points that are the waypoints.<br> We keep the raw data where depth was estimated between 0.2 m and 50.0 m deep. <br> The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available.<br> At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. <br> The size of the grid cells is 0.524 m. <br> The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <br> <h2> Generic folder structure </h2> YYYYMMDD_COUNTRYCODE-optionalplace_device_session-number <br> ├── DCIM : folder to store videos and photos depending on the media collected. <br> ├── GPS : folder to store any positioning related file. If any kind of correction is possible on files (e.g. Post-Processed Kinematic thanks to rinex data) then the distinction between device data and base data is made. If, on the other hand, only device position data are present and the files cannot be corrected by post-processing techniques (e.g. gpx files), then the distinction between base and device is not made and the files are placed directly at the root of the GPS folder. <br> │ ├── BASE : files coming from rtk station or any static positioning instrument. <br> │ └── DEVICE : files coming from the device. <br> ├── METADATA : folder with general information files about the session. <br> ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. <br> │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. <br> │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. <br> │ ├── IA : destination folder for image recognition predictions. <br> │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. <br> └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <br> <h2> Software </h2> All the raw data was processed using our <a href="https://github.com/SeatizenDOI/plancha-workflow/releases/tag/v1.0.3" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. <br>You can find all the necessary scripts to download this data in this <a href="https://github.com/SeatizenDOI/zenodo-tools" target="_blank">repository</a>. <br>Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>! <br>

Synthetic Dataset of charging processes by electric vehicles at workplace in Germany

<p>The dataset shows eight cluster groups that depict the mobility behavior of electric vehicle users in the employee context. For this purpose, 23.9 million data entries were analyzed, corresponding to 37,238 charging sessions. These data were collected over the year 2023. The 220 charging points were exclusively accessible to employees (private use case). From the data, cluster groups were derived using the Gaussian Mixture Model, and a synthetic dataset was generated through Monte Carlo sampling.</p> <p><span>The dataset consists of 8000 synthetic profiles, offering a robust scientific basis. By retaining the same statistical attributes as the empirical data, the synthetic profiles represent eight different mobility clusters, each containing 1000 entries, including full-time and part-time employees, shift workers, pool vehicle users, and opportunists.</span> Each cluster is represented by the mean parking start hours (arrival time - in decimal hours), mean parking duration (in decimal hours), the average energy recharged, and the average charging duration, each including the cluster-specific standard deviation and median.</p> <p>Further information can be obtained from the upcoming publication: "Synthetic Dataset of charging processes by electric vehicles at workplace in Germany."</p>

opencc-by-4.0Aug 2018View details →

Vehicle Routing Problem with Drones Instances

<p>The following instances are originally presented in the paper <em>An Adaptive Large Neighborhood Search Metaheuristic for the Vehicle Routing Problem with Drones</em>, written by David Sacramento Lechado, David Pisinger and Stefan Røpke, and published in <em>Transportation Research Part C: Emerging Technologies</em>.</p> <p>The data correspond to 112 instances corresponding to different scenarios for the Vehicle Routing Problem with Drones. Each instance is named <strong>n.m.t</strong>, where <strong>n</strong> is the number of customers in the scenario, <strong>m</strong> is the dimension of the grid, and <strong>t</strong> is the generic name of the scenario. Moreover, the data additionally contains 10 clustered instances, named <strong>n.m.c.t</strong>, where <strong>c</strong> refers to the cluster-feature of the instance, 100 instances for the sensitivity analysis, named <strong>n.m.s.t</strong>, where <strong>s</strong> stands for sensitivity, and 375 instances for the experiments with drone savings as function of grid size, named <strong>n.m.g.t</strong>.</p> <p>The first line of each instance file indicates the number of customers in the specific instance. The second line is the header for the characteristics of each customer, where it is written <em>Coordinate X, Coordinate Y</em> and <em>Demand. </em>The following provides the previous information for each customer in the instance.</p> <p>Finally, there is an extra file, named <strong>RouteData,</strong> which provides information about the value of the parameters in the main configuration of the problem. These parameters are:</p> <ul> <li><strong>TruckSpeed: </strong>Speed in mpm (miles per minute) of the trucks.</li> <li><strong>DroneSpeed: </strong>Speed in mpm of the drones.</li> <li><strong>TruckCapacityWithDrones: </strong>Maximum capacity of the trucks for the drone-truck scenario.</li> <li><strong>TruckCapacityWithoutDrones: </strong>Maximum capacity of the trucks for the truck-only scenario.</li> <li><strong>DroneCapacity:</strong> Maximum allowed weight a drone can carry.</li> <li><strong>ServiceTimeTruck:</strong> Required service time for a truck to service a customer.</li> <li><strong>ServiceTimeDrone:</strong> Required service time for a drone to service a customer.</li> <li><strong>Endurance: </strong>Maximum flight endurance of the battery of the drone.</li> <li><strong>MaximumDriveTime: </strong>Maximum duration time of the routes.</li> <li><strong>LaunchTime: </strong>Required time for launching a drone.</li> <li><strong>RecoveryTime: </strong>Required time for recovering a drone.</li> <li><strong>CostFactor:</strong> Corresponding parameter for computing the cost for a truck for traversing arc (i,j), given by fuel price (euro/liter), consumption rate (liter/km) and miles converter (km/miles).</li> <li><strong>DroneFactor: </strong>Corresponding parameter for computing the cost for a drone for traversing arc (i,j) with respect to the truck cost.</li> </ul> <p> </p>

opencc-by-4.0May 2019View details →

Vehicle CAN bus data

<p>The dataset contains 20Hz sampled CAN bus data from a passenger vehicle, e.g. WheelSpeed FL (speed of the front left wheel), SteerAngle (steering wheel angle), Role, Pitch, and accelerometer values per direction. Due to an export error, all GPS data is currently 0, but we are currently looking for a solution and will update the record as soon as possible.</p> <p> </p>

opencc-zeroSep 2024View details →

Data set for risk management in the allocation of vehicles to tasks in transport companies using a heuristic algorithm

<p>The purpose of this dataset is to enable the replication of the research results presented in the article: Izdebski, M. (2023). Risk management in the allocation of vehicles to tasks in transport companies using a heuristic algorithm. Archives of Transport, 67(3), 139-153. https://doi.org/10.5604/01.3001.0053.7463 - published online: 2023-09-30, which discusses the allocation problem of vehicles to tasks, taking into account risk issues.</p> <p>Dataset contains:</p> <ul> <li>Readme.txt: description of the dataset</li> <li>InputData.xlsx: Contains the input data used in the model</li> <li>DistributionFit.xlsx: Compliance testing and distribution parameters for road accidents of any type and collision-type</li> <li>OutputAssignment.xlsx: Results of assignment and alghoritm tests</li> </ul> <p>The dataset was created as part of the E-Laas project (Energy optimal urban logistics As A Service).<br>Project implemented as part of the call ERA-NET Cofund Urban Accessibility and Connectivity (ENUAC China Call) organized by JPI Urban Europe and the National Natural Science Foundation of China (NSFC). This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 875022.<br> E-Laas project is carried out in an international consortium. Project coordinator in Europe: Chalmers University of Technology (Sweden), project coordinator in China: Shanghai University (China), consortium members: Tsinghua University (China), Warsaw University of Technology (Poland), cooperation partners: Stockholms stad, Trafikkontoret (Sweden), ParkUnload (Spain), Metropolis GZM (Poland), Shanghai Urban-Rural Construction and Transportation Department (China), Volvo Group Trucks Technology and Operations (Sweden).<br>- The Chinese part of the project is funded by National Natural Science Foundation of China.<br>- The Swedish part of the project is funded by Swedish Energy Agency.<br>- The Polish part of the project is funded by the National Science Centre, Poland (project no. 2022/04/Y/ST8/00134). The value of the co-financing is PLN 878,107.00. Project duration 27/04/2023 - 26/04/2026 (36 months).</p>

Number of BEV (a battery electric vehicle) and PHEV (a plug-in hybrid electric vehicle) vehicles for each Country (2019)

<p>According to the Global E.V. Outlook 2020, China ranks first in vehicles in operation with electric or hybrid engines. In second place in the U.S. and third place in Norway.</p>

The monthly operating costs (vehicle, fuel, and maintenance) of each compared vehicle, Tesla 3 (283 HP), and Infiniti Q50 (300 HP).

<p>We compared two vehicles with similar horsepower, Tesla 3 (283 HP), and Infiniti Q50 (300 HP). The monthly operating costs of each compared vehicle were:</p> <ul> <li> <p>for the model, Tesla 3 electric vehicle was US$426.10/month, including purchase and depreciation US$333/month, fuel (electricity) US$27.8/month, maintenance US$65.3/month.</p> </li> <li> <p>for the model, Infiniti Q50, the internal combustion engine car was US$583.7/month, including purchase and depreciation US$321/month, fuel US$166/month, maintenance US$96.7/month.</p> </li> </ul>

Greenhouse gas emissions (lifecycle) of each compared vehicle, Tesla 3 (283 HP), and Infiniti Q50 (300 HP).

<p>We compared two vehicles with similar horsepower, Tesla 3 (283 HP), and Infiniti Q50 (300 HP). The CO_2 emissions for these vehicles  were: </p> <ul> <li> <p>for the model Tesla 3, CO_2 emissions were 161.8 gCO_2 eq/mile, including 31.8 gCO_2 eq/mile in vehicle production, 25 gCO_2 eq/mile in battery production, and 105 gCO_2 eq/mile in electricity production.</p> </li> <li> <p>for the model Infiniti Q50, CO_2 emissions were 503.8 gCO_2 eq/mile , including vehicle production 40.5 gCO_2 eq/mile, fuel production 91.3 gCO_2 eq/mile, in-service combustion 372 gCO_2 eq/mile.</p> </li> </ul>

Vehicle Trips Dataset without Missing Values

<p>This dataset contains 329 daily time series representing the number of trips and vehicles belonging to a set of for-hire vehicle (FHV) companies. The dataset was originally extracted from https://github.com/fivethirtyeight/uber-tlc-foil-response.</p> <p>The original dataset contains missing values and they have been replaced by carrying forward the corresponding last seen observations (LOCF method).</p>

Vehicle Trips Dataset with Missing Values

Underwater images collected by an Autonomous Surface Vehicle in St-Leu, Réunion - 2023-11-10

<i>This dataset was collected by an Autonomous Surface Vehicle in St-Leu, Réunion - 2023-11-10.</i> <br> <br><br>Underwater or aerial images collected by scientists or citizens can have a wide variety of use for science, management, or conservation. These images can be annotated and shared to train IA models which can in turn predict the objects on the images. We provide a set of tools (hardware and software) to collect marine data, predict species or habitat, and provide maps.<br><br> This dataset is part of larger collection referencing numerous underwater and aerial images <a href="https://doi.org/10.5281/zenodo.11125847" target="_blank">Seatizen Altas</a>. Methods, tools and scientific objectives are also described in a dedicated data paper.<br> <h2>Image acquisition</h2> This session has 24.43 GB of MP4 files, which were trimmed into 8257 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 100.0% of these extracted images are useful and 0.0% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. <br> Multilabel predictions have been made on useful frames using <a href="https://huggingface.co/lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze" target="_blank">DinoVd'eau</a> model. <br> <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. <br> Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 91.58 %, Q2: 7.65 %, Q5: 0.77 % <br> <h2> Bathymetry </h2> The data are collected using a single-beam echosounder <a href="https://www.echologger.com/products/single-frequency-echosounder-deep" target="_blank">ETC 400</a>. <br> We only keep the values which have a GPS correction in Q1.<br> We keep the points that are the waypoints.<br> We keep the raw data where depth was estimated between 0.2 m and 50.0 m deep. <br> The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available.<br> At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. <br> The size of the grid cells is 0.299 m. <br> The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <br> <h2>Photogrammetry</h2> OpenDroneMap software was used to create an orthophoto from the raw images. <br> Here is the list of parameters different from the default values for the orthophoto generation. <br> For more details, you can read the log.json file or the 000_photogrammatry_report.pdf report. <br><br> <code> {'auto_boundary': True, 'cog': True, 'fast_orthophoto': True, 'feature_quality': 'ultra', 'gps_accuracy': 0.1, 'max_concurrency': 50, 'optimize_disk_space': True, 'orthophoto_resolution': 0.1, 'rolling_shutter': True, 'skip_3dmodel': True} </code> <h2> Generic folder structure </h2> YYYYMMDD_COUNTRYCODE-optionalplace_device_session-number <br> ├── DCIM : folder to store videos and photos depending on the media collected. <br> ├── GPS : folder to store any positioning related file. If any kind of correction is possible on files (e.g. Post-Processed Kinematic thanks to rinex data) then the distinction between device data and base data is made. If, on the other hand, only device position data are present and the files cannot be corrected by post-processing techniques (e.g. gpx files), then the distinction between base and device is not made and the files are placed directly at the root of the GPS folder. <br> │ ├── BASE : files coming from rtk station or any static positioning instrument. <br> │ └── DEVICE : files coming from the device. <br> ├── METADATA : folder with general information files about the session. <br> ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. <br> │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. <br> │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. <br> │ ├── IA : destination folder for image recognition predictions. <br> │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. <br> └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <br> <h2> Software </h2> All the raw data was processed using our <a href="https://github.com/SeatizenDOI/plancha-workflow/releases/tag/v1.0.3" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. <br>You can find all the necessary scripts to download this data in this <a href="https://github.com/SeatizenDOI/zenodo-tools" target="_blank">repository</a>. <br>Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>! <br>

Underwater images collected by an Autonomous Surface Vehicle in St-Leu, Réunion - 2023-11-10

<i>This dataset was collected by an Autonomous Surface Vehicle in St-Leu, Réunion - 2023-11-10.</i> <br> <br><br>Underwater or aerial images collected by scientists or citizens can have a wide variety of use for science, management, or conservation. These images can be annotated and shared to train IA models which can in turn predict the objects on the images. We provide a set of tools (hardware and software) to collect marine data, predict species or habitat, and provide maps.<br><br> This dataset is part of larger collection referencing numerous underwater and aerial images <a href="https://doi.org/10.5281/zenodo.11125847" target="_blank">Seatizen Altas</a>. Methods, tools and scientific objectives are also described in a dedicated data paper.<br> <h2>Image acquisition</h2> This session has 34.89 GB of MP4 files, which were trimmed into 10570 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 100.0% of these extracted images are useful and 0.0% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. <br> Multilabel predictions have been made on useful frames using <a href="https://huggingface.co/lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze" target="_blank">DinoVd'eau</a> model. <br> <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. <br> Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 69.18 %, Q2: 25.61 %, Q5: 5.22 % <br> <h2> Bathymetry </h2> The data are collected using a single-beam echosounder <a href="https://ceruleansonar.com/products/sounder-s500" target="_blank">S500</a>. <br> We only keep the values which have a GPS correction in Q1.<br> We keep the points that are the waypoints.<br> We keep the raw data where depth was estimated between 0.2 m and 50.0 m deep. <br> The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available.<br> At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. <br> The size of the grid cells is 0.195 m. <br> The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <br> <h2>Photogrammetry</h2> OpenDroneMap software was used to create an orthophoto from the raw images. <br> Here is the list of parameters different from the default values for the orthophoto generation. <br> For more details, you can read the log.json file or the 000_photogrammatry_report.pdf report. <br><br> <code> {'auto_boundary': True, 'cog': True, 'fast_orthophoto': True, 'feature_quality': 'ultra', 'gps_accuracy': 0.1, 'max_concurrency': 50, 'optimize_disk_space': True, 'orthophoto_resolution': 0.1, 'rolling_shutter': True, 'skip_3dmodel': True} </code> <h2> Generic folder structure </h2> YYYYMMDD_COUNTRYCODE-optionalplace_device_session-number <br> ├── DCIM : folder to store videos and photos depending on the media collected. <br> ├── GPS : folder to store any positioning related file. If any kind of correction is possible on files (e.g. Post-Processed Kinematic thanks to rinex data) then the distinction between device data and base data is made. If, on the other hand, only device position data are present and the files cannot be corrected by post-processing techniques (e.g. gpx files), then the distinction between base and device is not made and the files are placed directly at the root of the GPS folder. <br> │ ├── BASE : files coming from rtk station or any static positioning instrument. <br> │ └── DEVICE : files coming from the device. <br> ├── METADATA : folder with general information files about the session. <br> ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. <br> │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. <br> │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. <br> │ ├── IA : destination folder for image recognition predictions. <br> │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. <br> └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <br> <h2> Software </h2> All the raw data was processed using our <a href="https://github.com/SeatizenDOI/plancha-workflow/releases/tag/v1.0.3" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. <br>You can find all the necessary scripts to download this data in this <a href="https://github.com/SeatizenDOI/zenodo-tools" target="_blank">repository</a>. <br>Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>! <br>

Underwater images collected by an Autonomous Surface Vehicle in Boucan, Réunion - 2023-11-09

<i>This dataset was collected by an Autonomous Surface Vehicle in Boucan, Réunion - 2023-11-09.</i> <br> <br><br>Underwater or aerial images collected by scientists or citizens can have a wide variety of use for science, management, or conservation. These images can be annotated and shared to train IA models which can in turn predict the objects on the images. We provide a set of tools (hardware and software) to collect marine data, predict species or habitat, and provide maps.<br><br> This dataset is part of larger collection referencing numerous underwater and aerial images <a href="https://doi.org/10.5281/zenodo.11125847" target="_blank">Seatizen Altas</a>. Methods, tools and scientific objectives are also described in a dedicated data paper.<br> <h2>Image acquisition</h2> This session has 14.54 GB of MP4 files, which were trimmed into 4660 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 100.0% of these extracted images are useful and 0.0% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. <br> Multilabel predictions have been made on useful frames using <a href="https://huggingface.co/lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze" target="_blank">DinoVd'eau</a> model. <br> <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. <br> Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 7.63 %, Q2: 91.59 %, Q5: 0.78 % <br> <h2>Photogrammetry</h2> OpenDroneMap software was used to create an orthophoto from the raw images. <br> Here is the list of parameters different from the default values for the orthophoto generation. <br> For more details, you can read the log.json file or the 000_photogrammatry_report.pdf report. <br><br> <code> {'auto_boundary': True, 'cog': True, 'fast_orthophoto': True, 'feature_quality': 'ultra', 'gps_accuracy': 0.1, 'max_concurrency': 44, 'optimize_disk_space': True, 'orthophoto_resolution': 0.1, 'rolling_shutter': True, 'skip_3dmodel': True} </code> <h2> Generic folder structure </h2> YYYYMMDD_COUNTRYCODE-optionalplace_device_session-number <br> ├── DCIM : folder to store videos and photos depending on the media collected. <br> ├── GPS : folder to store any positioning related file. If any kind of correction is possible on files (e.g. Post-Processed Kinematic thanks to rinex data) then the distinction between device data and base data is made. If, on the other hand, only device position data are present and the files cannot be corrected by post-processing techniques (e.g. gpx files), then the distinction between base and device is not made and the files are placed directly at the root of the GPS folder. <br> │ ├── BASE : files coming from rtk station or any static positioning instrument. <br> │ └── DEVICE : files coming from the device. <br> ├── METADATA : folder with general information files about the session. <br> ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. <br> │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. <br> │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. <br> │ ├── IA : destination folder for image recognition predictions. <br> │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. <br> └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <br> <h2> Software </h2> All the raw data was processed using our <a href="https://github.com/SeatizenDOI/plancha-workflow/releases/tag/v1.0.3" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. <br>You can find all the necessary scripts to download this data in this <a href="https://github.com/SeatizenDOI/zenodo-tools" target="_blank">repository</a>. <br>Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>! <br>

Underwater images collected by an Autonomous Surface Vehicle in St-Leu, Réunion - 2024-03-13

<i>This dataset was collected by an Autonomous Surface Vehicle in St-Leu, Réunion - 2024-03-13.</i> <br> <br><br>Underwater or aerial images collected by scientists or citizens can have a wide variety of use for science, management, or conservation. These images can be annotated and shared to train IA models which can in turn predict the objects on the images. We provide a set of tools (hardware and software) to collect marine data, predict species or habitat, and provide maps.<br><br> This dataset is part of larger collection referencing numerous underwater and aerial images <a href="https://doi.org/10.5281/zenodo.11125847" target="_blank">Seatizen Altas</a>. Methods, tools and scientific objectives are also described in a dedicated data paper.<br> <h2>Image acquisition</h2> This session has 21.93 GB of MP4 files, which were trimmed into 6033 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 99.9% of these extracted images are useful and 0.1% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. <br> Multilabel predictions have been made on useful frames using <a href="https://huggingface.co/lombardata/DinoVdeau-large-2024_04_03-with_data_aug_batch-size32_epochs150_freeze" target="_blank">DinoVd'eau</a> model. <br> <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. <br> Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 94.64 %, Q2: 5.34 %, Q5: 0.02 % <br> <h2> Bathymetry </h2> The data are collected using a single-beam echosounder <a href="https://www.echologger.com/products/single-frequency-echosounder-deep" target="_blank">ETC 400</a>. <br> We only keep the values which have a GPS correction in Q1.<br> We keep the points that are the waypoints.<br> We keep the raw data where depth was estimated between 0.2 m and 10.0 m deep. <br> The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available.<br> At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. <br> The size of the grid cells is 0.111 m. <br> The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <br> <h2>Photogrammetry</h2> OpenDroneMap software was used to create an orthophoto from the raw images. <br> Here is the list of parameters different from the default values for the orthophoto generation. <br> For more details, you can read the log.json file or the 000_photogrammatry_report.pdf report. <br><br> <code> {'auto_boundary': True, 'cog': True, 'fast_orthophoto': True, 'feature_quality': 'ultra', 'gps_accuracy': 0.1, 'max_concurrency': 50, 'optimize_disk_space': True, 'orthophoto_resolution': 0.1, 'rolling_shutter': True, 'skip_3dmodel': True} </code> <h2> Generic folder structure </h2> YYYYMMDD_COUNTRYCODE-optionalplace_device_session-number <br> ├── DCIM : folder to store videos and photos depending on the media collected. <br> ├── GPS : folder to store any positioning related file. If any kind of correction is possible on files (e.g. Post-Processed Kinematic thanks to rinex data) then the distinction between device data and base data is made. If, on the other hand, only device position data are present and the files cannot be corrected by post-processing techniques (e.g. gpx files), then the distinction between base and device is not made and the files are placed directly at the root of the GPS folder. <br> │ ├── BASE : files coming from rtk station or any static positioning instrument. <br> │ └── DEVICE : files coming from the device. <br> ├── METADATA : folder with general information files about the session. <br> ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. <br> │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. <br> │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. <br> │ ├── IA : destination folder for image recognition predictions. <br> │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. <br> └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <br> <h2> Software </h2> All the raw data was processed using our <a href="https://github.com/SeatizenDOI/plancha-workflow/releases/tag/v1.0.3" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. <br>You can find all the necessary scripts to download this data in this <a href="https://github.com/SeatizenDOI/zenodo-tools" target="_blank">repository</a>. <br>Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>! <br>