Find research datasets worth reusing

opencc-by-4.0Sep 2024View details →

Synthetic dataset: Traffic Accidents & Vehicle Registry

<ul> <li>Dataset A: Traffic Accidents -- Comprising 15,000 records with attributes such as Accident ID, Date, Time, Location, Vehicle ID, Severity, and Description.</li> <li>Dataset B: Vehicle Registry -- Comprising 20,000 records with attributes including Vehicle ID, Owner Age, Owner Gender, Vehicle Model, and Registration Date.</li> </ul>

opencc-by-4.0Sep 2024View details →

Synthetic Fleet Generation and Vehicle Assignment to Synthetic Households for Regional and Sub-regional Sustainability Analysis

This dataset provides the MOVES-Matrix emission and energy use rates for the NCST project "Synthetic Fleet Generation and Vehicle Assignment to Synthetic Households for Regional and Sub-regional Sustainability Analysis" by the Georgia Tech research team.   The abstract of the project is as follows. In this study, a modeling framework was developed to generate high-resolution synthetic fleets, for use with synthetic household modeling in activity-based travel models, by integrating various data sources. The synthetic households were generated by pairing household locations and demographic attributes, and synthetic fleets were assigned to the households so that travel demand model outputs would have vehicles associated with each model-predicted tour for energy and emissions analysis. The CO emissions were modeled for each vehicle and each link traversed by vehicles as predicted by the travel demand model, and the results of the synthetic fleet (by employing Monte Carlo simulations and Bootstrap techniques) were compared with those from standard regional and sub-regional fleet configurations. The results demonstrated that using a traditional sub-regional fleet scenario produced 30% higher predicted emissions than when the synthetic fleet was employed with predicted vehicle trips, and that using a regional average fleet (applied throughout the region) produced emissions that were more than 50% higher than synthetic fleet emissions. Lowest household emissions were associated with low-income and non-working households, and highest emissions were associated with moderate-income households and one-person high-income household groups. The results presented in the research are not necessarily conclusive, because the licensed vehicle data procured for Atlanta appear to be biased toward older vehicles. Model year penetration rates are accounted for in these analyses, but the authors believe that the variability in the registration mix for newer vehicles is likely underestimated in the data procured for these analyses. The authors conclude that access to statewide registration data will be required to remove potential biases that exist in licensed private data sets. Nevertheless, the study does demonstrate that properly pairing vehicle model years with the most active households (and their daily trips) significantly impacts energy and emissions analysis.

opencc-by-4.0Oct 2024View details →

Measurement report: Unexpected high volatile organic compounds emission from vehicles on the Tibetan Plateau Dataset

This dataset includes various emission profiles and related data, specifically: <ol> <li> Source Profile Data at Different Altitudes. </li> <li> Emission Factor Data. </li> <li> Emission Ratio Data. </li> <li>Source Profile Data from PMF Source Apportionment: Data obtained through Positive Matrix Factorization (PMF), revealing the composition of emission sources.</li> <li> Average Profiles of Gasoline Vapors: Derived from sealed housing evaporative determination (SHED) tests, with references 1-7. </li> <li> Average Profiles of Gasoline Vehicle Exhaust: Based on dynamometer tests, with references 2, 8-13. </li> <li> Average Profiles of Vehicular Emissions: Collected from low-altitude tunnel measurements, reflecting emissions in real-world driving scenarios, with references 14-24. </li> </ol>

opencc-by-4.0Nov 2024View details →

Enhanced Vehicle Detection via YOLOv7-Tiny, dataset we used

Our improvement on YOLOv7-tiny, with Sim_DFC, Inner-shape IoU, and BiFPN. These are the datasets we used.

opencc-by-4.0Nov 2023View details →

Trends in global fuel economy of new vehicles: 2005 - 2022

This data contains information on the fuel economy of new light-duty vehicle sales in major automotive markets and underpins the report - Trends in the global vehicle fleet 2023: managing the SUV shift and the EV transition. This report is the 2023 installment of the Global Fuel Economy Initiative (https://www.globalfueleconomy.org/) which, among other things, tracks progress in the efficiency of the global vehicle fleet.  Yearly vehicle sales are defined by segment (small car, medium car, large car, small SUV, large SUV, Light Commercial Vehicle) and powertrain ( internal combustion engine, mild hybrid, hybrid, plug-in hybrid, battery electric, fuel cell hydrogen). For each combination of segment and powertrain the data contains information on sales, average weight (kg), average footprint (m2), and average specific energy consumption (lge/100 km). The data was obtained from a set of sources and processed to gain the best possible estimate of global trends in energy consumption for light duty vehicles. More information on the methodologies underpinning the data analysis can be found in the report. Files included: <ul> <li>supplementary_information_GFEI2023_TDC.xlsx <ul> <li>data: sheet containing the the data -  blank rows indicate missing data.</li> <li>unique values: contains the unique values taken by each column in sheet data</li> </ul> </li> <li> supplementary_information_GFEI2023_vizualisation.xlsx <ul> <li>raw_data: same data as that available in the file above</li> <li>country_map: tha mapping of individual countries to regions used in the report</li> <li>filter_data: combinations of powertrain-year-country that account for less than 0.5% of yearly sales that are not displayed in the Graphs tab</li> <li>pivots_country, pivots_powertrain, pivots_segment: intermediate step </li> <li>Powertrain: registrations, specific fuel consumption, weight and footprint  by powertrain</li> <li>Segment: registrations, specific fuel consumption, weight and footprint  by segment</li> <li>Graphs: Includes plots of the above two tabs, contains a country selector in I2</li> </ul> </li> <li>sdmx.zip <ul> <li>Data structures and data in SDMX-ML format, containing exactly the same information as supplementary_information_GFEI2023_TDC.xlsx, converted using the code at <a href="https://github.com/transport-data/gfei-2023" target="_blank" rel="noopener">https://github.com/transport-data/gfei-2023</a></li> </ul> </li> </ul> Nota Bene: Older versions of Excel might encounter issues with titles in graphs, that is likely caused by the CONCAT function not being availble - substituting the CONCAT function with CONCATENATE should solve the problem.

opencc-by-4.0Nov 2024View details →

Spatially Gridded Vehicle Count Analysis for Connaught Place, New Delhi

Our Gridded Vehicle Count Analysis for Connaught Place, New Delhi, is a detailed geospatial analysis focusing on vehicular emissions in a highly trafficked urban area. Here's an overview: <h3>Objective</h3> The analysis aims to quantify vehicle density in Connaught Place, a major commercial and tourist hub, by generating a spatially gridded vehicle count dataset. By doing this, the project provides insights into traffic patterns and potential emission hotspots, supporting air quality forecasts and urban planning. <h3>Data Collection</h3> <ol> <li>Satellite Imagery: You collected high-resolution WorldView satellite imagery, zoomed specifically over Connaught Place to ensure the detail needed for vehicle detection.</li> <li>Image Preprocessing: Using PyQGIS, you loaded the satellite imagery, created a 0.09-degree buffer around Connaught Place, and generated a grid layer with 150m x 150m cells to spatially segment the area for analysis.</li> <li>Geotagged .tiff Files: The images were exported as geotagged .tiff files, preserving spatial information for each grid cell, facilitating precise location-based analysis.</li> </ol> <h3>Detection and Analysis</h3> <ol> <li>Object Detection Models: You employed YOLOv8 through YOLOv10 and other state-of-the-art deep learning models to detect various vehicle types, such as cars, buses, and trucks. This approach helped accurately identify vehicle counts in real time.</li> <li>Class-wise Detection: Specific object classes were defined, allowing for detailed counts by vehicle type, which is crucial for emission factor calculations.</li> <li>Result Export: The detected vehicle counts and corresponding latitude/longitude data were stored in netCDF files, enabling the creation of gridded emission inventories.</li> </ol> <h3>Outcomes and Applications</h3> The gridded dataset enables real-time monitoring of vehicular density and emissions over time. This data: <ul> <li>Enhances air quality models by integrating localized emission sources.</li> <li>Assists policymakers in targeting emission reduction initiatives.</li> <li>Provides a foundational layer for studies on urban traffic flow and its environmental impacts in high-density zones like Connaught Place.</li> </ul> In summary, this analysis provides a scientifically robust and spatially precise view of vehicular density, supporting both atmospheric studies and practical urban management decisions.

opencc-by-4.0Jul 2021View details →

Energy consumption of 15 electric vehicles (one day resolution)

Energy consumption of 15 electric vehicles (one day resolution) Sérgio Ramos, João Soares, Zahra Foroozandeh, Inês Tavares, Zita Vale Paper title: TODO Type: EV consumption Duration: One year Resolution: One day Application: Paper submitted on Sheets description: <ul> <li>EV 1-15: Contains the information of the energy consumption and initial State of Charge of each EV (kWh).</li> </ul>

opencc-by-4.0Sep 2021View details →

Constrained Fitness Landscape Analysis of Vehicle Routing Problems

The repository is a set of Jupyter notebooks and datasets used in the article titled: “Constrained Fitness Landscape Analysis for Capacitated Vehicle routing problems.” The main objective is to give the tools for reproducing the results shown in the paper.

Underwater images collected by an Autonomous Surface Vehicle in Boucan, Réunion - 2023-09-07

This dataset was collected by an Autonomous Surface Vehicle in Boucan, Réunion - 2023-09-07. 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. 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. <h2>Image acquisition</h2> This session has 78.22 GB of MP4 files, which were trimmed into 17724 frames (at 2997/1000 fps). The frames are georeferenced. 29.87% of these extracted images are useful and 70.13% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. 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. <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. Device GPS : Emlid Reach M2 Quality of our data - Q1: 45.47 %, Q2: 8.05 %, Q5: 46.48 % <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>. We only keep the values which have a GPS correction in Q1. We keep the points that are the waypoints. We keep the raw data where depth was estimated between 0.2 m and 50.0 m deep. The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available. At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. The size of the grid cells is 0.68 m. The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <h2>Photogrammetry</h2> OpenDroneMap software was used to create an orthophoto from the raw images. Here is the list of parameters different from the default values for the orthophoto generation. For more details, you can read the log.json file or the 000_photogrammatry_report.pdf report. <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 ├── DCIM : folder to store videos and photos depending on the media collected. ├── 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. │ ├── BASE : files coming from rtk station or any static positioning instrument. │ └── DEVICE : files coming from the device. ├── METADATA : folder with general information files about the session. ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. │ ├── IA : destination folder for image recognition predictions. │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <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>. All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. 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>. Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>!

Underwater images collected by an Autonomous Surface Vehicle in Cap-Homard, Réunion - 2023-11-28

This dataset was collected by an Autonomous Surface Vehicle in Cap-Homard, Réunion - 2023-11-28. 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. 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. <h2>Image acquisition</h2> This session has 36.25 GB of MP4 files, which were trimmed into 8316 frames (at 2997/1000 fps). The frames are georeferenced. 99.77% of these extracted images are useful and 0.23% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. 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. <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. Device GPS : Emlid Reach M2 Quality of our data - Q1: 18.63 %, Q2: 80.96 %, Q5: 0.41 % <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>. We only keep the values which have a GPS correction in Q1. We keep the points that are the waypoints. We keep the raw data where depth was estimated between 0.2 m and 50.0 m deep. The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available. At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. The size of the grid cells is 0.408 m. The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <h2>Photogrammetry</h2> OpenDroneMap software was used to create an orthophoto from the raw images. Here is the list of parameters different from the default values for the orthophoto generation. For more details, you can read the log.json file or the 000_photogrammatry_report.pdf report. <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 ├── DCIM : folder to store videos and photos depending on the media collected. ├── 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. │ ├── BASE : files coming from rtk station or any static positioning instrument. │ └── DEVICE : files coming from the device. ├── METADATA : folder with general information files about the session. ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. │ ├── IA : destination folder for image recognition predictions. │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <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>. All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. 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>. Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>!

Underwater images collected by an Autonomous Surface Vehicle in Boucan, Réunion - 2023-11-22

This dataset was collected by an Autonomous Surface Vehicle in Boucan, Réunion - 2023-11-22. 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. 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. <h2>Image acquisition</h2> This session has 35.76 GB of MP4 files, which were trimmed into 9704 frames (at 2997/1000 fps). The frames are georeferenced. 99.99% of these extracted images are useful and 0.01% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. 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. <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. Device GPS : Emlid Reach M2 Quality of our data - Q1: 95.98 %, Q2: 3.96 %, Q5: 0.06 % <h2> Bathymetry </h2> The data are collected using a single-beam echosounder <a href="https://ceruleansonar.com/products/sounder-s500" target="_blank">S500</a>. We only keep the values which have a GPS correction in Q1. We keep the points that are the waypoints. We keep the raw data where depth was estimated between 0.2 m and 50.0 m deep. The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available. At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. The size of the grid cells is 0.451 m. The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <h2>Photogrammetry</h2> OpenDroneMap software was used to create an orthophoto from the raw images. Here is the list of parameters different from the default values for the orthophoto generation. For more details, you can read the log.json file or the 000_photogrammatry_report.pdf report. <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 ├── DCIM : folder to store videos and photos depending on the media collected. ├── 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. │ ├── BASE : files coming from rtk station or any static positioning instrument. │ └── DEVICE : files coming from the device. ├── METADATA : folder with general information files about the session. ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. │ ├── IA : destination folder for image recognition predictions. │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <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>. All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. 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>. Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>!

Underwater images collected by an Autonomous Surface Vehicle in Boucan, Réunion - 2023-11-21

This dataset was collected by an Autonomous Surface Vehicle in Boucan, Réunion - 2023-11-21. 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. 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. <h2>Image acquisition</h2> This session has 21.24 GB of MP4 files, which were trimmed into 7790 frames (at 2997/1000 fps). The frames are georeferenced. 98.45% of these extracted images are useful and 1.55% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. 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. <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. Device GPS : Emlid Reach M2 Quality of our data - Q1: 33.31 %, Q2: 65.47 %, Q5: 1.22 % <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>. We only keep the values which have a GPS correction in Q1. We keep the points that are the waypoints. We keep the raw data where depth was estimated between 0.2 m and 50.0 m deep. The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available. At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. The size of the grid cells is 0.346 m. The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <h2>Photogrammetry</h2> OpenDroneMap software was used to create an orthophoto from the raw images. Here is the list of parameters different from the default values for the orthophoto generation. For more details, you can read the log.json file or the 000_photogrammatry_report.pdf report. <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 ├── DCIM : folder to store videos and photos depending on the media collected. ├── 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. │ ├── BASE : files coming from rtk station or any static positioning instrument. │ └── DEVICE : files coming from the device. ├── METADATA : folder with general information files about the session. ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. │ ├── IA : destination folder for image recognition predictions. │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <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>. All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. 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>. Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>!

opencc-by-4.0May 2024View details →

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

This dataset was collected by an Autonomous Surface Vehicle in St-Leu, Réunion - 2023-11-10. 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. 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. <h2>Image acquisition</h2> This session has 22.68 GB of MP4 files, which were trimmed into 7786 frames (at 2997/1000 fps). The frames are georeferenced. 99.99% of these extracted images are useful and 0.01% are useless, according to predictions made by <a href="jacques-v0.1.0_model-20240513_v20.0" target="_blank">Jacques model</a>. 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. <h2> GPS information: </h2> The data was processed with a PPK workflow to achieve centimeter-level GPS accuracy. Base : Files coming from rtk a GPS-fixed station or any static positioning instrument which can provide with correction frames. Device GPS : Emlid Reach M2 Quality of our data - Q1: 84.7 %, Q2: 15.17 %, Q5: 0.13 % <h2> Bathymetry </h2> The data are collected using a single-beam echosounder <a href="https://ceruleansonar.com/products/sounder-s500" target="_blank">S500</a>. We only keep the values which have a GPS correction in Q1. We keep the points that are the waypoints. We keep the raw data where depth was estimated between 0.2 m and 50.0 m deep. The data are first referenced against the WGS84 ellipsoid. Then we apply the local geoid if available. At the end of processing, the data are projected into a homogeneous grid to create a raster and a shapefiles. The size of the grid cells is 0.208 m. The raster and shapefiles are generated by linear interpolation. The 3D reconstruction algorithm is ballpivot. <h2>Photogrammetry</h2> OpenDroneMap software was used to create an orthophoto from the raw images. Here is the list of parameters different from the default values for the orthophoto generation. For more details, you can read the log.json file or the 000_photogrammatry_report.pdf report. <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 ├── DCIM : folder to store videos and photos depending on the media collected. ├── 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. │ ├── BASE : files coming from rtk station or any static positioning instrument. │ └── DEVICE : files coming from the device. ├── METADATA : folder with general information files about the session. ├── PROCESSED_DATA : contain all the folders needed to store the results of the data processing of the current session. │ ├── BATHY : output folder for bathymetry raw data extracted from mission logs. │ ├── FRAMES : output folder for georeferenced frames extracted from DCIM videos. │ ├── IA : destination folder for image recognition predictions. │ └── PHOTOGRAMMETRY : destination folder for reconstructed models in photogrammetry. └── SENSORS : folder to store files coming from other sources (bathymetry data from the echosounder, log file from the autopilot, mission plan etc.). <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>. All predictions were generated by our <a href="https://github.com/SeatizenDOI/plancha-inference/releases/tag/v1.0.0" target="_blank">inference pipeline</a>. 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>. Enjoy your data with <a href="https://github.com/SeatizenDOI" target="_blank">SeatizenDOI</a>!

dryad36/100

Source Data for Crowdsourcing Bridge Dynamic Monitoring with Smartphone Vehicle Trips

This data accompanies the study "Crowdsourcing Bridge Dynamic Monitoring with Smartphone Vehicle Trips" published in (Nature) Communications Engineering. This paper focuses on using large and inexpsensive datasets for obtaining information on the dynamics of bridges. In this study, data is collected by smartphones in moving vehicles as the cross over a bridge, in three distinct applications. Smartphone data was collected in controlled field experiments and uncontrolled Uber rides on a long-span suspension bridge in the USA (The Golden Gate Bridge) and an analytical method was developed to accurately recover modal properties. The method was also successfully applied to partially-controlled crowdsourced data collected on a short-span highway bridge in Italy. The results suggest that larve and inexpensive datasets collected by smartphones could play a role in monitoring the health of existing transportation infrastructure. The data provided includes the source data for the figures in the publication as well as the "controlled data" referenced in the study.

opencc-zeroNov 2022View details →

dryad36/100

Vehicle pollution is associated with elevated insect damage to street trees

1. Vehicle pollution is a pervasive aspect of anthropogenic change across rural and urban habitats. The most common emissions are carbon- or nitrogen-based pollutants that may impact diverse interactions between plants and insect herbivores. However, the effects of vehicle pollution on plant-insect interactions are poorly understood. 2. Here, we combine a city-wide experiment across the Sacramento Metropolitan Area and a laboratory experiment to determine how vehicle emissions affect insect herbivory and leaf nutritional quality. 3. We demonstrate that leaf damage to a native oak species (Quercus lobata) commonly planted across the western US is substantially elevated on trees exposed to vehicle emissions. In the laboratory, caterpillars preferred leaves from highway-adjacent trees and performed better on leaves from those same trees. 4. Synthesis and applications. Together, our studies demonstrate that the heterogeneity in vehicle emissions across cities may explain highly variable patterns of insect herbivory on street trees. Our results also indicate that trees next to highways are particularly vulnerable to multiple stressors, including insect damage. To combat these effects, urban foresters may consider planting trees that are less susceptible to insect herbivory along heavily traveled roadways.

opencc-zeroNov 2022View details →

Grey model analysis of vehicle population, road transport energy consumption, and vehicular emissions

The files provide additional information to the paper “Grey model analysis of vehicle population, road transport energy consumption, and vehicular emissions". The supporting data file contains excel sheets of data used in the analysis, and the supporting information file provides some assumptions, background information and other results not included in the paper

opencc-zeroNov 2022View details →

dryad36/100

Data from: Multiobjective optimization algorithm for accurate MADYMO reconstruction of vehicle-pedestrian accidents

Uncertainty in reconstruction accuracy is a critical problem faced in the current traffic accident reconstruction process. The purpose of this study is to explore the use of an improved optimization algorithm combined with MAthematical DYnamic MOdels (MADYMO) multibody simulations and crash data to conduct accurate reconstructions of vehicle–pedestrian accidents. The performance of three commonly employed multiobjective optimization algorithms, including nondominated sorting genetic algorithm-II (NSGA-II), neighbourhood cultivation genetic algorithm (NCGA) and multiobjective particle swarm optimization (MOPSO) were compared and evaluated. The effects of the number of objective functions, the selection of different objective functions and the optimal number of iterations are also investigated. The present study indicated that NSGA-II had better convergence and generated more noninferior solutions and better final solutions than NCGA and MOPSO. And multibody simulations coupled with optimization algorithms can be used to accurately reconstruct vehicle-pedestrian collisions.

Interactive Widget – Acceptance of autonomous vehicles dataset exploration

<a href="https://research-data.shinyapps.io/CAV_Acceptance/">https://research-data.shinyapps.io/CAV_Acceptance/</a> The following widget gives citizens access to data collected with the Connected and Autonomous Vehicle Acceptance Assessment Tool (CAVA) developed in the PAsCAL project (Public acceptance of Connected and Autonomous vehicles). The aim of the CAVA is to measure autonomous vehicle acceptance via evaluation of expected autonomous vehicle consequences. A survey was employed with over 5000 participants from 11 countries.

KR, Vehicle quality of service support, Tethering via vehicle using mmWave communication

Use Case Category: Vehicle quality of service support User Story: Tethering via vehicle using mmWave communication Location: Korean (KR) trial site According to 3GPP TS 22.186 R16, Vehicle quality of service support “enables a V2X application to be timely notified of expected or estimated change of quality of service before actual change occurs and to enable the 3GPP System to modify the quality of service in line with V2X application’s quality of service needs. Based on the quality of service information, the V2X application can adapt behaviour to 3GPP System’s conditions. The benefits of this use case group are offerings of smoother user experience of service”. User Story: Tethering via vehicle using mmWave communication Tethering via Vehicle use case enables in-vehicle UEs and pedestrian UEs to access the network with the help of a vehicle relay which is deployed at a vehicle. For in-vehicle UEs, through Tethering via Vehicle use case, it is possible to avoid high penetration loss occurring from the metallic vehicle surface, thereby achieving more reliable wireless connectivity as well as reduced UE power consumption. The in-vehicle UEs are also benefited from the minimized handover operations. Only the vehicle relays involve in the handover operations. For pedestrian UEs, tethering via Vehicle use case enables more reliable connectivity, increased throughput and reduced UE power consumption since it reduces the communication range of the pedestrian UEs. As a deployment scenario for Tethering via Vehicle use case, a vehicle relay can have the Internet connectivity to the network through a base station (BS) including macrocell BS, microcell BS, and BS-type road side unit (RSU). Another UE such as UE-type RSU can also provide the Internet connectivity to the network. Non-terrestrial links such as satellite BS, satellite relay, and high-altitude platform (HAP) can also be used to provide the Internet connectivity to the network. Tethering via Vehicle use case generally supports eMBB-type services such as web surfing, FTP, and video streaming. Hence, it intrinsically requires high data throughput up to several Gbps. In order to satisfy such very high throughput, large bandwidth is necessary which is quite difficult in lower frequency bands below 6 GHz. Therefore, mmWave frequency band should be employed to support such high throughput and to satisfy the Tethering via Vehicle use case.

CN, Vehicles Platooning, Cloud-assisted platooning

Use Case Category: Vehicles Platooning User Story: Cloud-assisted platooning Location: Chinese (CN) trial site According to 3GPP TS 22.186 R16, Vehicles Platooning “enables the vehicles to dynamically form a group travelling together. All the vehicles in the platoon receive periodic data from the leading vehicle, in order to carry on platoon operations. This information allows the distance between vehicles to become extremely small, i.e., the gap distance translated to time can be very low (sub second). Platooning applications may allow the vehicles following to be autonomously driven”. User Story: Cloud-assisted platooning The autonomous driving vehicle fleet communicates with each other through LTE-V at the start. Among them, the leading vehicle includes the platoon control unit (PCU), which coordinates the vehicles in the fleet to ensure a certain safe distance and to drive in a platoon. The leading vehicle communicates with the control centre deployed in a cloud server through V2N to obtain the test scheme and the global path planning. Then it provides the basic planning for the rear vehicle through V2V communication (including chasing, continuous running, acceleration, deceleration, obstacle avoidance, overall acceleration, and deceleration, etc.). The following vehicle also has a certain perception and planning decision-making ability. Besides, LTE-V communication can be replaced by DSRC technology, and comparison between these two methods will be implemented.