---------------- EXample 1 ----------------
READ ME File For 'Dataset in support of the thesis Predicting high-resolution terrain properties in robotic survey applications'

Dataset DOI: 10.5258/SOTON/D3357

ReadMe Author: Jose Cappelletto, University of Southampton

This dataset contains the experimental data used to generate the figures included in the main text of the thesis "Predicting high-resolution terrain properties in robotic survey applications", more specifically supporting section "4.1 Predicted habitat maps"

The dataset contain 4 CSV files, each one corresponding to a different experiment. Each file contains georeferenced data corresponding to the predicted seafloor habitat classes present. The experiment-file are named as follows,

    pred_mean_10m_geoclr_L15m_h16_5955:      Best contrastive learning model (geoCLR)
    pred_mean_20m_ae_L45m_h128_0722.csv:     Best Location Guided Autoencoder (LGA)
    pred_near_30m_geoclr_L45m_h04_3824.csv:  Worst contrastive learning model (geoCLR)
    pred_near_30m_ae_L45m_h04_1300.csv:      Worst Location Guided Autoencoder (LGA)

Each one of the 4 (FOUR) CSV files include the same headers and are explained below;

    index: Sequential identifier for each row in the dataset (INTEGER).
    UUID: Universally unique identifier for distinguishing each entry (XXX_YYYY format).
    easting [m]: Eastward coordinate in meters in projected coordinate system (EPSG:32629).
    northing [m]: Northward coordinate in meters in projected coordinate system (EPSG:32629).
    longitude [deg]: Geographical longitude in degrees for spatial reference (EPSG:4326).
    latitude [deg]: Geographical latitude in degrees for spatial reference (EPSG:4326).
    pred_vector_class_0: Predicted probability value for class 0, RIPPLE.
    pred_vector_class_1: Predicted probability value for class 1, BEDROCK.
    pred_vector_class_2: Predicted probability value for class 2, SEDIMENT.
    pred_vector_class_3: Predicted probability value for class 3, BOULDER.
    std_vector_class_0: Standard deviation of predictions for class 0, RIPPLE.
    std_vector_class_1: Standard deviation of predictions for class 1, BEDROCK.
    std_vector_class_2: Standard deviation of predictions for class 2, SEDIMENT.
    std_vector_class_3: Standard deviation of predictions for class 3, BOULDER.
    max_class: Index of the class with the highest predicted probability for the entry.

Date of data collection: July 2022

Date of data processing: August 2023

Information about geographic location of data collection: Haig Fras Marine Protected Area, U.K.

Credits:
    - Adrian Bodenmann, University of Southampton, U.K., adrian.bodenmann@soton.ac.uk
    - Catherine Wardell, National Oceanographic Centre, Southampton, U.K., catherine.wardell@noc.ac.uk

Related projects:
    -   PhD project: Predicting high-resolution terrain properties in robotic survey applications

Dataset available under a CC BY NC 4.0 licence

Date that the file was created: January 2025.