This paper introduces a dataset (called STships) based on real seascapes for ship detection. The dataset was collected with a HD dome camera in Sanya on July 18 and July 20, 2022. 7483 images are covered in the dataset belonging to 6 types of targets (research vessel, oil tanker, general cargo ship, fishing boat, passenger ship and ocean buoy). 7109 images are carefully annotated, in addition to 374 background images. A number of practical influencing factors were taken into account during the construction of the dataset, including: background, visible proportion, scale, illumination. Baseline detector is trained on the dataset to evaluate the performance. And there is also a comparison between STships and other dataset. The dataset is expected to advance ship detection and application under real seascapes


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    STships: An Object Detection Dataset Based on Real Seascapes


    Contributors:
    Wang, Jiahao (author) / Shi, Yang (author)


    Publication date :

    2023-09-23


    Size :

    777744 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Seascapes

    Fleming, Douglas K. | Taylor & Francis Verlag | 1982


    Rock Art and Seascapes

    McNiven, Ian J. / Brady, Liam M. | Wiley | 2012


    REVIEWS - Ships and Seascapes

    Cordingly | Online Contents | 1998


    Bentley, Bridenthal and Wigen (eds): Seascapes

    Cadigan, Sean T. | Online Contents | 2009


    UAV Embedded Real-Time Object Detection by a DCNN Model Trained on Synthetic Dataset

    Bernardo, Ricardo Maroquio / Claudio Batista da Silva, Luis / Ferreira Rosa, Paulo Fernando | IEEE | 2023