Find research datasets worth reusing
Search datasets from major research repositories and use ShareScore to quickly assess how well each record supports discovery, access, and reuse.
622
datasets available to search
ShareScore release 0.9.0
Dataset results
622 results for “underwater”
Ten years (2013-2023) of fish assemblage data collected seasonally with underwater visual surveys on paired artificial and natural reefs
<p>The study of assembly patterns and dynamics of organisms has long remained a foundational theme in ecology. Further, the relationship between assemblages and different habitats can provide important insight on ecological processes and guide management and conservation efforts (e.g., restoration, protected areas). We conducted underwater visual surveys of reef fish assemblages at 14 sites in the eastern Gulf of Mexico, including eight that were paired artificial and natural reefs. By using a paired design, we controlled biotic (e.g., larval supply), abiotic (e.g., depth), and socio variables (e.g., fishing access) to isolate the effect of reef type. Trained scientific SCUBA divers with extensive experience with reef fishes from the broader tropical western Atlantic region conducted two to four 10-minute stationary surveys on the paired reefs each season (i.e., calendar quarters) for 10 years from spring 2013 to spring 2023. We also surveyed six additional artificial reefs from winter 2020 to spring 2023 that lacked natural reef pairs. During each survey, the divers identified and estimated the total lengths of all taxa<strong> </strong>observed within an imaginary cylinder around them. The imaginary cylinders had a radius up to 7.5 meters (depending on horizontal visibility) and extended from the seafloor to the highest visible water above the diver. During the period of study, we conducted a total of 1,349 surveys and counted 544,736 fish that represented 171 taxa (most at the species level). Analyses of these data have revealed habitat-specific heterogeneity of the fish assemblages at both taxonomic and functional trait levels, the importance of herbivory in structuring the benthos, and socio-ecological interactions in the system, among other findings. These data may be useful for other researchers interested in patterns and dynamics of populations and communities, functional traits, taxa-habitat relationships, and for parameterizing statistical, joint distribution, metacommunity, and ecosystem models. In addition, because many of the observed taxa<strong> </strong>are of management concern, they may be useful for researchers interested in fisheries science. The data are free to use, are not copyright restricted, and we ask users to cite this data paper.</p>
R-CAUSTIC: Rippling CAUSTICs underwater Image dataset
<p><strong>Description</strong></p><p>Rippling caustics seem to be the main factor degrading the underwater RGB image quality and affecting the image- based 3D reconstruction process in very shallow waters. These effects are adversely affecting image matching algorithms by throwing off most of them, leading to less accurate matches and causing issues in the Simultaneous Localization and Mapping (SLAM) based navigation of the Remotely Operated Vehicles (ROV) and Autonomous Underwater Vehicles (AUV) on shallow waters. Also, they are the main cause for dissimilarities in the generated textures and orthoimages. In order to fill the gap in the literature regading underwater rippling caustics imagery with real ground truth and reference images, the first real-world underwater caustics benchmark dataset which contains 1465 underwater images is presented. Together with the RGB imagery, the corresponding generated ground truth images are delivered for facilitating the training and testing of machine learning and deep learning methods for image classification. R-CAUSTIC dataset also provides the necessary data to evaluate, at least to some extent, the performance of 3D reconstruction approaches. Data were acquired using a GoPro Hero 4 Black action camera with image dimensions of 4000 x 3000 pixels, focal length of 2.77mm and pixel size of 1.55μm and a tripod. Action cameras are widely used for underwater image acquisition. The dataset was captured in near-shore underwater sites at depths varying from 0.5 to 2m. No artificial light sources were used. Due to the wind, the turbulent surface of the water created dynamic rippling caustics on the seabed. In total 1465 RGB images were collected, separated in 7 different datasets; five of them containing stereo images, one of them tri-stereo images and one consists of multi-stereo imagery acquired in 7 different camera poses.</p><p> </p><p><strong>Publication</strong></p><p>The paper is availbale in Open Access here: https://ieeexplore.ieee.org/document/10172291</p><p><strong>If you use this dataset please cite it as R-CAUSTIC</strong> [Reference].<br>[Reference]: <strong>P. Agrafiotis, K. Karantzalos and A. Georgopoulos, "Seafloor-Invariant Caustics Removal From Underwater Imagery," in </strong><i><strong>IEEE Journal of Oceanic Engineering</strong></i><strong>, vol. 48, no. 4, pp. 1300-1321, Oct. 2023, doi: 10.1109/JOE.2023.3277168.</strong></p><p>BibTeX:</p><p>@ARTICLE{10172291, author={Agrafiotis, Panagiotis and Karantzalos, Konstantinos and Georgopoulos, Andreas}, journal={IEEE Journal of Oceanic Engineering}, title={Seafloor-Invariant Caustics Removal From Underwater Imagery}, year={2023}, volume={48}, number={4}, pages={1300-1321}, doi={10.1109/JOE.2023.3277168}}</p><p> </p>
Supporting Data for the paper titled "The Intensity, Directionality and Statistics of Underwater Noise from Melting Icebergs"
<p>The dataset contains:</p> <p>a) 13 audio files, with names including date, track number and channel; format: WAV files</p> <p>b) data from magnetic compass used to calculate noise directionality; format: txt files with lines containing date, time and magnetic direction (degrees)</p> <p>c) GPS tracks of the boat and attached acoustic buoy; format: txt files with NMEA codes</p> <p>d) GPS tracks around each iceberg tracked; format: txt files with UTM coordinates</p> <p>The study was founded by National Science Centre Poland grant no. 2013/11/N/ST10/01729 and partially supported within statutory activities No 3841/E-41/S/2018 of the Ministry of Science and Higher Education of Poland. Partial support for this work was also provided by US Office of Naval Research, Grant No. N00014-17-1-2633.</p> <p>Corresponding author: Oskar Glowacki, oglowacki@igf.edu.pl</p>
An Acoustic and Optical Dataset for the Perception of Underwater Unexploded Ordnance (UXO)
<p>We present a dataset for acoustic and optical sensing of unexploded ordnance (UXO) underwater.</p> <p>UXO in the sea pose an environmental problem and a challenge for the growing offshore economy. It is best practice to perform the recovery of ammunition without explosions to protect anthropogenic structures and marine mammals. During explosive ordnance disposal (EOD), experts often rely on optical images. However, visibility underwater may be limited in harbor areas, after storm events or in waters with very mobile sediments. Thus, visual inspection is not always possible. EOD experts therefore use high-frequency sonars with large vertical apertures like the ARIS Explorer 3000 for acoustic imaging. While efforts have been made to use the available information for 3D reconstruction, existing solutions can be limited to predefined motion patterns. </p> <p>The topic is inherently sensitive, and most of the data is acquired by and for private companies and not made available to the public, which impedes research in this area. Additionally, in-situ data often lacks sufficient pose information. To facilitate further research, we created a validation dataset that was recorded in a controlled experimental environment. It has the following properties:</p> <ul> <li>Close to 100 recordings of 3 different UXO.</li> <li>More than 74000 matched and annotated imaging sonar and camera frames.</li> <li>UXO ground truths in the form of photogrammetric 3D models.</li> <li>Precise position and attitude sensor data with respect to the targets.</li> <li>Realistic motion trajectories achievable in non-experimental environments.</li> </ul> <p>This dataset allows quantitative analysis with different algorithms. 3D models and trajectories can be compared against each other to evaluate different solutions.</p> <p> </p> <p><strong>The accompanying paper is:</strong></p> <blockquote> <p>@INPROCEEDINGS{dahn2024uxo,<br> author={Dahn, Nikolas and Firvida, Miguel Bande and Sharma, Proneet and Christensen, Leif and Geisle, Oliver and Mohrmann, Jochen and Frey, Torsten and Kumar Sanghamreddy, Prithvi and Kirchner, Frank},<br> booktitle={OCEANS 2024 - Halifax}, <br> title={An Acoustic and Optical Dataset for the Perception of Underwater Unexploded Ordnance (UXO)}, <br> year={2024},<br> doi={10.1109/OCEANS55160.2024.10754316}}</p> </blockquote> <p>The paper is available on <a href="https://www.researchgate.net/publication/386124306_An_Acoustic_and_Optical_Dataset_for_the_Perception_of_Underwater_Unexploded_Ordnance_UXO">researchgate</a>.</p> <p> </p> <p><strong>Notes:</strong></p> <ul> <li>Labels have been (unfortunately) generated for the SD camera frames. To get the correct coordinates on the included FHD images, multiply all coordinates by 3.</li> </ul> <p> </p> <p><strong>Files</strong>:</p> <ul> <li>data_export_recordings.7z: main dataset</li> <li>data_export_polar.7z: contains only the polar-transformed sonar frames</li> <li>data_export_3dmodels.7z: 3d models of the UXO</li> <li>data_processed.7z: extracted and cut unmatched raw data</li> </ul>
Nephrops (Nephrops norvegicus) Burrow object detection simple training dataset from Irish Underwater TV surveys
<div> <div> <div> <div> <h1>Training dataset</h1> <p>Norway prawns (<em>Nephrops norvegicus</em>), also known as the Dublin Bay prawn, are common around the Irish coast. They are found in distinct sandy/muddy areas where the sediment is suitable for them to construct their burrows. <em>Nephrops </em>spend a great deal of time in their burrows and their emergence from these is related to time of year, light intensity and tidal strength. The Irish <em>Nephrops </em>fishery is extremely valuable with landings recently worth around €55m at first sale, supporting an important Irish fishing industry. </p> <p><em>Nephrops</em> are managed in Functional Units (FUs). The Marine Institute has conducted under water television surveys since 2002 to independently estimate abundance, distribution and stock sizes of <em>Nephrops</em> <em>norvegicus </em>for:</p> <ul> <li>Irish Sea <em>Nephrops</em> Grounds (FU 14 and 15) in collaboration with <a title="Link to 'Fisheries and Aquatic Ecosystems' work in AFBI Northern Ireland" href="https://www.afbini.gov.uk/area-of-expertise/fisheries-and-aquatic-ecosystems">AFBI</a> an <a title="Link to Cefas (the Centre for Environment, Fisheries, and Aquaculture Science) in the UK" href="https://www.cefas.co.uk/">CEFAS</a>.</li> <li>Porcupine Bank <em>Nephrops</em> Grounds (FU16)</li> <li>Aran, Galway Bay and Slyne Head <em>Nephrops</em> Grounds (FU17)</li> <li>South and South west Ireland <em>Nephrops</em> Grounds (FU19)</li> <li>Labadie, Jones and Cockburn <em>Nephrops</em> Grounds (FU20 and 21)</li> <li>“Smalls” <em>Nephrops</em> Grounds (FU22)</li> </ul> <p>Each year during the summer months, on average 300 stations are surveyed each year, in three survey legs, covering all the FUs in depths from 20 to 650 metres.</p> <p>A high definition camera system is towed over the sea bed for 10 minutes travelling approx. 200m at 0.8 knots on a purpose built sledge. The UWTV survey follows survey protocols available <a title="Link to survey protocols" href="https://doi.org/10.17895/ices.pub.8014">here</a> agreed by International Council for the Exploration of the Sea (ICES) Working Group on <em>Nephrops </em>surveys (WGNEPS). </p> <p>As part of the iMagine project a selection of images from the Underwater TV survey Functional Units were annotated with bounding boxes and labels in YOLOv8 format to train an YOLOv8 Object Detection Models. The training dataset is saved in YOLOv8 format. It is intended to train a YOLOv8 Nephrrops burrow object detection model to assess the utility of an Object Detection model is assisting Prawn Survey work in the semi automated annotation of prawn burrow imagery.</p> </div> </div> </div> </div>
The VAROS Synthetic Underwater Data Set: Towards realistic multi-sensor underwater data with ground truth
<p>Underwater visual perception requires being able to deal with bad and rapidly varying illumination and with reduced visibility due to water turbidity. The verification of such algorithms is crucial for safe and efficient underwater exploration and intervention operations. Ground truth data play an important role in evaluating vision algorithms. However, obtaining ground truth from real underwater environments is in general very hard, if possible at all. In a synthetic underwater 3D environment, however, (nearly) all parameters are known and controllable, and ground truth data can be absolutely accurate in terms of geometry. In this paper, we present the VAROS environment, our approach to generating highly realistic underwater video and auxiliary sensor data with precise ground truth, built around the Blender modeling and rendering environment. VAROS allows for physically realistic motion of the simulated underwater (UW) vehicle including moving illumination. Pose sequences are created by first defining way-points for the simulated underwater vehicle which are expanded into a smooth vehicle course sampled at IMU data rate (200Hz). This expansion uses a vehicle dynamics model and a discrete-time controller algorithm that simulates the sequential following of the way-points. The scenes are rendered using the raytracing method, which generates realistic images, integrating direct light, and indirect volumetric scattering. The VAROS dataset version 1 provides images, inertial measurement unit (IMU) and depth gauge data, as well as ground truth poses, depth images and surface normal images.</p>
SOUNDSCAPE North Adriatic Underwater Noise Sound Pressure Levels
<p>Within the Interreg Italy-Croatia <strong>SOUNDSCAPE</strong> <strong>project</strong> a basin-scale, cross-national, long-term underwater monitoring in the Northern Adriatic Sea was carried out (https://www.italy-croatia.eu/web/soundscape). A broad network of nine monitoring stations, characterized by different natural conditions and anthropogenic pressures, ensured acoustic data collection from March 2020 to June 2021, including the first full lockdown period related to the COVID-19 pandemic (March–April 2020). Calibrated stationary recorders featured with an omnidirectional Neptune Sonar D60 Hydrophone recorded continuously 24h a day (48 kHz, 16 bit).</p> <p>Data were analysed to Sound Pressure Levels (SPLs, dB re 1 uPa) that are here released as a dataset composed of 20 and 60 seconds averaged SPL output files for each station. Data are archived using structured hdf5 files, each one containing metadata and SPL data, according to ICES (International Council for the Exploration of the Sea) continuous noise data specification (https://www.ices.dk/data/data-portals/Pages/Continuous-Noise.aspx).</p> <p>A research object with a jupyter notebook developed to post process SPL data is available at https://doi.org/10.24424/hrhm-8849</p> <p>If data are used, please cite also Petrizzo, A., Barbanti, A., Barfucci, G. <em>et al.</em> Author Correction: First assessment of underwater sound levels in the Northern Adriatic Sea at the basin scale. <em>Sci Data</em> <strong>10</strong>, 179 (2023). https://doi.org/10.1038/s41597-023-02099-x</p>
Underwater sounds, including killer whale and humpback whale vocalizations, recorded in northern Norway in January 2023
<p>Dataset of underwater acoustic recordings obtained during the expedition “Orcalize” that took place in Skjervøy in northern Norway from 29<sup>th</sup> December 2022 till 6<sup>th</sup> January 2023. The data contains vocalizations from killer whales and songs from humpback whales which gather in the local fjords during the winter months to feed on herring. We recorded in the band of 20 Hz – 60 kHz with calibrated hydrophones arranged in a compact tetrahedral array that we deployed over board of a motorboat. In total we provide 16 files of continuous recordings with duration from several minutes to over one hour. The total dataset is about 7 hours 37 minutes long and the memory size is 62.8 GB. See the file info.pdf for more information.</p>
Test experiments with distributed acoustic sensing and hydrophone arrays for locating underwater sounds.
<p>Whales and dolphins rely on sound for navigation and communication, making them an intriguing subject for studying language evolution. Traditional hydrophone arrays have been used to record their acoustic behavior, but optical fibers have emerged as a promising alternative. This study explores the use of distributed acoustic sensing (DAS), a technique that detects local stress in optical fibers, for underwater sound recording. An experiment was conducted in Lake Zurich, where a fiber-optic cable and a self-made hydrophone array were deployed. A test signal was broadcasted at various locations, and the resulting data was synchronized and consolidated into files. Analysis revealed distinct frequency responses in the DAS channels and provided insights into sound propagation in the lake. Challenges related to cable sensitivity, sample rate, and broadcast fidelity were identified. This dataset serves as a valuable resource for advancing acoustic sensing techniques in underwater environments, especially for studying marine mammal vocal behavior.</p>
Underwater surveys of mullet schools (Mugil liza) with Adaptive Resolution Imaging Sonar
<p>This dataset is part of a research project that employs deep learning, with a density-based regression approach, to count fish in low-resolution sonar images (Tarling et al. preprint arXiv DOI: http://arxiv.org/abs/2104.14964).</p> <p>In this repository, we provide data from sonar-based underwater videos of schools of migratory mullets (<em>Mugil liza</em>) recorded at the Tesoura beach (28.495775 S, 48.759996 W), a 100-meter long beach at the inlet canal connecting the Laguna lagoon system to the Atlantic Ocean, in southern Brazil. Since the water transparency at the lagoon canal is very low (from 0.3 to 1.5m visibility; collected <em>in situ</em> with a Secchi disk), mullet schools were recorded by deploying an Adaptive Resolution Imaging Sonar, ARIS 3000 (Sound Metrics Corp, WA, USA), which uses 128 beams to project a wedge-shaped volume of acoustic energy and convert their returning echoes into a digital overhead view of the mullet schools.</p> <p>This dataset contains 500 fully annotated images that were manually marked for the location and abundance of mullet fish, and 126 raw sonar video files, representing over 100k images. The files are organized as follows:</p> <p>1) "2018-MM-DD_HHMMSS" files are mp4 videos (you may need to add the file extension ".mp4"): There are 126 ARIS files converted into MP4 videos totalling over 789MB of underwater footage captured at 3 frames/seconds. Note that file names indicate the date and time the video was recorded.</p> <p>2) ".npy" files (in Labelled_data.zip): From the video files, 500 images were selected for labelling. Images (x) were cropped to represent a 4x8.5m<sup>2</sup> area and resized to 320 x 576 pixels. Mullet fish were marked with a point annotation. Corresponding ground truth density maps (y) were generated by convolving a Gaussian kernel over the image mask, size =4 and standard deviation = 1. The labelled dataset was randomly split into a holdout partition of 350 training images, 70 validation, and 80 test. </p> <p>3) ".csv" files: log of frames selected for the labelled subset of data</p> <p>4) ".h5" file: pre-trained weights for our multi-task with uncertainty regularisation network</p> <p>To advance the development of these machine learning tools, we also make our code openly available (https://github.com/ptarling/DeepLearningFishCounting).</p>
Underwater temperature, light, and dissolved oxygen data from 3 mini-buoys in Lake Sunapee, NH, USA from June to October 2018
Three mini-buoys were deployed during a portion of the ice-off period of 2018 in Lake Sunapee, NH, USA with HOBO temperature sensors at various depths below the water’s surface. Temperature data were collected using HOBO pendant temperature and HOBO pendant temperature/light sensors at descending depths between 0.1m and the nearest whole and/or half meter increments below the water surface and above the sediment/water interface at 10 minute intervals. Two buoys (Georges Mills and Herrick Cove) also had miniDOT (PME) dissolved oxygen and temperature sensors placed 1.75 meters below the surface. The buoys were located in cove areas of the lake in the north east arm of the lake (Herrick Cove, 0.1m – 6.5m), the west side of the southern area of the lake near Lake Sunapee State Beach (State Beach, 0.1m – 2.5m) and the northwest arm of the lake (George’s Mills, 0.1m -7m). This data has been QAQC’d to remove obviously errant data and artifacts of buoy maintenance visits.
Underwater Photosynthetically Active Radiation (PAR) from in-situ Lake Primary Production Experiments in the McMurdo Dry Valleys of Antarctica (1995-2024, ongoing)
As part of the Long Term Ecological Research (LTER) project in the McMurdo Dry Valleys of Antarctica, PAR is measured at one depth during primary production experiments in Dry Valley Lakes. This data set quantifies instantaneous underwater PAR at 10-meter depths in Lakes Bonney and Hoare, and at 7-meter depth in Lake Fryxell (depths vary after 2008, check data file for actual depth). Ambient PAR was also measured at the air-ice interface from each of these lakes.
Underwater photosynthetically active radiation (PAR) vertical profiles collected from lakes in the McMurdo Dry Valleys, Antarctica (1993-2025, ongoing)
As part of the Long Term Ecological Research (LTER) project in the McMurdo Dry Valleys of Antarctica, lakes are monitored for photosynthetic active radiation (PAR) levels. This data set exemplifies the vertical profile of underwater PAR profiles in perennial ice covered lakes and ambient PAR measured at the air-ice interface. Profile data can be used to calculate extinction coefficients for the water column and ice layer.
Palmer Deep Underwater Acoustic Mooring Deployments 2021-2024
Marine mammals play a crucial role in the ocean ecosystem, yet monitoring their occurrence, distribution, and abundance poses a challenge in remote areas, such as the Antarctic. Traditionally, human observers have conducted visual surveys to detect marine mammals, relying on the animals' need to surface periodically for air. This method is often costly, requiring a large team of observers and the use of ships or aircraft. Additionally, visual surveys are limited by weather and sighting conditions, such as fog, rain, heavy seas, and darkness. Despite their expense, visual surveys are often inefficient for continuous real-time monitoring of marine mammal presence. However, they remain essential for tasks like photo identification, health assessment, and abundance estimation for many species. In recent decades, passive acoustic recorders have become extremely popular for detecting vocally active marine mammals, as they can operate continuously for periods of months to years. This enables a passive method of observation of marine mammals during periods where visual surveys are untenable. We developed a long term mooring system equipped with an underwater acoustic recorder that is capable of recording at 60% for up to 200 days at a time. Since 2022, we have been maintaining these moorings to provide for annual acoustic coverage, and will continue to under the Palmer LTER. This data can be used to compose the underwater soundscape of the region, and a general assessment of year-round acoustic presence of sound producing animals. Further analysis could examine the impacts of inter- and intra- seasonal environmental conditions (e.g., sea ice advance and retreat, sea ice extent, storminess) on the phenology of acoustic presence in the region.
Underwater images collected by Scuba diving in Ifaty, Madagascar - 2023-05-07
<i>This dataset was collected by Scuba diving in Ifaty, Madagascar - 2023-05-07.</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 2.66 GB of MP4 files, which were trimmed into 1056 frames (at 2997/1000 fps). <br> The frames are not georeferenced. <br> 93.75% of these extracted images are useful and 6.25% 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> No GPS. <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://doi.org/10.5281/zenodo.15853010" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://doi.org/10.5281/zenodo.15228535" 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 Scuba diving in Nosy-Ve, Madagascar - 2023-04-30
<i>This dataset was collected by Scuba diving in Nosy-Ve, Madagascar - 2023-04-30.</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 1.5 GB of JPG files, which were trimmed into 267 frames (at 4 fps). <br> The frames are georeferenced. <br> 99.63% of these extracted images are useful and 0.37% 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> No GPS. <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://doi.org/10.5281/zenodo.15853010" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://doi.org/10.5281/zenodo.15228535" 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 Sarodrano, Madagascar - 2023-05-03
<i>This dataset was collected by an Autonomous Surface Vehicle in Sarodrano, Madagascar - 2023-05-03.</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 15.07 GB of MP4 files, which were trimmed into 5136 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 59.54% of these extracted images are useful and 40.46% 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: 0.0 %, Q2: 0.0 %, Q5: 100.0 % <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://doi.org/10.5281/zenodo.15853010" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://doi.org/10.5281/zenodo.15228535" 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 Ifaty, Madagascar - 2023-05-06
<i>This dataset was collected by an Autonomous Surface Vehicle in Ifaty, Madagascar - 2023-05-06.</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 17.33 GB of MP4 files, which were trimmed into 4251 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> Base : No Base <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 0.0 %, Q2: 0.0 %, Q5: 100.0 % <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 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. <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 0.422 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://doi.org/10.5281/zenodo.15853010" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://doi.org/10.5281/zenodo.15228535" 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 Anakao, Madagascar - 2023-05-01
<i>This dataset was collected by an Autonomous Surface Vehicle in Anakao, Madagascar - 2023-05-01.</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 13.58 GB of MP4 files, which were trimmed into 2767 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 80.88% of these extracted images are useful and 19.12% 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: 0.0 %, Q2: 0.0 %, Q5: 100.0 % <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 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. <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 0.335 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://doi.org/10.5281/zenodo.15853010" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://doi.org/10.5281/zenodo.15228535" 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 Ifaty, Madagascar - 2023-05-07
<i>This dataset was collected by an Autonomous Surface Vehicle in Ifaty, Madagascar - 2023-05-07.</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 25.86 GB of MP4 files, which were trimmed into 5425 frames (at 2997/1000 fps). <br> The frames are georeferenced. <br> 99.17% of these extracted images are useful and 0.83% 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> Base : No Base <br> Device GPS : Emlid Reach M2 <br> Quality of our data - Q1: 0.0 %, Q2: 0.0 %, Q5: 100.0 % <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 keep the points that are the waypoints.<br> We keep the raw data where depth was estimated between 0.2 m and 20.0 m deep. <br> The data are first referenced against the WGS84 ellipsoid. <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 0.313 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://doi.org/10.5281/zenodo.15853010" target="_blank">worflow</a>. <br>All predictions were generated by our <a href="https://doi.org/10.5281/zenodo.15228535" 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>
ScienceDex guides
Understand access before you commit
These curated guides explain access requirements, typical timelines, costs, and reuse considerations for widely used research datasets.
Allen Brain Atlas
Allen Brain Atlas is an Allen Institute collection of brain map atlases, datasets, APIs, and analysis tools covering mouse, human, and non-human primate brain resources.
Annotated Behaviour and Observability Dataset (ABODe)
ABODe is a University of Edinburgh DataShare dataset for behavior classification in group-housed mice using home-cage video, identities, bounding boxes, ground-plate positions, and annotator labels.
DANDI Archive for NWB datasets
DANDI is a BRAIN Initiative archive for publishing and sharing neurophysiology data, including electrophysiology, optophysiology, and behavioral data packaged as NWB and related standards.
International Brain Laboratory public data
The International Brain Laboratory public data releases expose standardized mouse decision-making experiments, including Neuropixels recordings, widefield calcium imaging, behavior, and session metadata accessed through the ONE API.
OpenNeuro
OpenNeuro is a free, open platform for sharing neuroimaging datasets, with public search, dataset pages, and download paths for web, S3, DataLad, and the OpenNeuro CLI.