Deepcloud - A Fully Convolutionnal Neural Network for Cloud and Shadow Masking in Optical Satellite Images

    Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

    Abstract

    Many cloud and shadow detection methods have been proposed already, but improvements can be made on accuracy or automation. In this study, we propose a Fully Convolutional Network model for the detection of clouds and shadows in optical satellite images. The proposed model was trained on 165 Landsat images in Finland, and tested on an independent set of images. The cloud and shadow detection accuracy reached 95%, outperforming both quantitatively and qualitatively a selection of other deep learning architectures.
    Original languageEnglish
    Title of host publication2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings
    PublisherIEEE Institute of Electrical and Electronic Engineers
    Pages2107-2110
    Number of pages4
    ISBN (Electronic)978-1-5386-7150-4 , 978-1-5386-7149-8
    ISBN (Print)978-1-5386-7151-1
    DOIs
    Publication statusPublished - 5 Nov 2018
    MoE publication typeNot Eligible
    EventIEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Valencia, Spain
    Duration: 22 Jul 201827 Jul 2018

    Conference

    ConferenceIEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018
    Abbreviated titleIGARSS
    CountrySpain
    CityValencia
    Period22/07/1827/07/18

    Fingerprint

    detection method
    automation
    Landsat
    learning
    satellite image
    detection

    Keywords

    • Remote sensing
    • Clouds
    • Satellites
    • Artificial satellites
    • Earth
    • Optical imaging
    • Cloud and shadow masking
    • optical images
    • deep learning
    • fully convolutional network
    • Landsat

    Cite this

    Molinier, M., Reunanen, N., Lämsä, A., Astola, H., & Räty, T. (2018). Deepcloud - A Fully Convolutionnal Neural Network for Cloud and Shadow Masking in Optical Satellite Images. In 2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings (pp. 2107-2110). [8517484] IEEE Institute of Electrical and Electronic Engineers . https://doi.org/10.1109/IGARSS.2018.8517484
    Molinier, Matthieu ; Reunanen, Niko ; Lämsä, Arttu ; Astola, Heikki ; Räty, Tomi. / Deepcloud - A Fully Convolutionnal Neural Network for Cloud and Shadow Masking in Optical Satellite Images. 2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings. IEEE Institute of Electrical and Electronic Engineers , 2018. pp. 2107-2110
    @inproceedings{5e438abdd62d4fc18091324372dc11ac,
    title = "Deepcloud - A Fully Convolutionnal Neural Network for Cloud and Shadow Masking in Optical Satellite Images",
    abstract = "Many cloud and shadow detection methods have been proposed already, but improvements can be made on accuracy or automation. In this study, we propose a Fully Convolutional Network model for the detection of clouds and shadows in optical satellite images. The proposed model was trained on 165 Landsat images in Finland, and tested on an independent set of images. The cloud and shadow detection accuracy reached 95{\%}, outperforming both quantitatively and qualitatively a selection of other deep learning architectures.",
    keywords = "Remote sensing, Clouds, Satellites, Artificial satellites, Earth, Optical imaging, Cloud and shadow masking, optical images, deep learning, fully convolutional network, Landsat",
    author = "Matthieu Molinier and Niko Reunanen and Arttu L{\"a}ms{\"a} and Heikki Astola and Tomi R{\"a}ty",
    year = "2018",
    month = "11",
    day = "5",
    doi = "10.1109/IGARSS.2018.8517484",
    language = "English",
    isbn = "978-1-5386-7151-1",
    pages = "2107--2110",
    booktitle = "2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings",
    publisher = "IEEE Institute of Electrical and Electronic Engineers",
    address = "United States",

    }

    Molinier, M, Reunanen, N, Lämsä, A, Astola, H & Räty, T 2018, Deepcloud - A Fully Convolutionnal Neural Network for Cloud and Shadow Masking in Optical Satellite Images. in 2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings., 8517484, IEEE Institute of Electrical and Electronic Engineers , pp. 2107-2110, IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018, Valencia, Spain, 22/07/18. https://doi.org/10.1109/IGARSS.2018.8517484

    Deepcloud - A Fully Convolutionnal Neural Network for Cloud and Shadow Masking in Optical Satellite Images. / Molinier, Matthieu; Reunanen, Niko; Lämsä, Arttu; Astola, Heikki; Räty, Tomi.

    2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings. IEEE Institute of Electrical and Electronic Engineers , 2018. p. 2107-2110 8517484.

    Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

    TY - GEN

    T1 - Deepcloud - A Fully Convolutionnal Neural Network for Cloud and Shadow Masking in Optical Satellite Images

    AU - Molinier, Matthieu

    AU - Reunanen, Niko

    AU - Lämsä, Arttu

    AU - Astola, Heikki

    AU - Räty, Tomi

    PY - 2018/11/5

    Y1 - 2018/11/5

    N2 - Many cloud and shadow detection methods have been proposed already, but improvements can be made on accuracy or automation. In this study, we propose a Fully Convolutional Network model for the detection of clouds and shadows in optical satellite images. The proposed model was trained on 165 Landsat images in Finland, and tested on an independent set of images. The cloud and shadow detection accuracy reached 95%, outperforming both quantitatively and qualitatively a selection of other deep learning architectures.

    AB - Many cloud and shadow detection methods have been proposed already, but improvements can be made on accuracy or automation. In this study, we propose a Fully Convolutional Network model for the detection of clouds and shadows in optical satellite images. The proposed model was trained on 165 Landsat images in Finland, and tested on an independent set of images. The cloud and shadow detection accuracy reached 95%, outperforming both quantitatively and qualitatively a selection of other deep learning architectures.

    KW - Remote sensing

    KW - Clouds

    KW - Satellites

    KW - Artificial satellites

    KW - Earth

    KW - Optical imaging

    KW - Cloud and shadow masking

    KW - optical images

    KW - deep learning

    KW - fully convolutional network

    KW - Landsat

    UR - http://www.scopus.com/inward/record.url?scp=85063150877&partnerID=8YFLogxK

    U2 - 10.1109/IGARSS.2018.8517484

    DO - 10.1109/IGARSS.2018.8517484

    M3 - Conference article in proceedings

    SN - 978-1-5386-7151-1

    SP - 2107

    EP - 2110

    BT - 2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings

    PB - IEEE Institute of Electrical and Electronic Engineers

    ER -

    Molinier M, Reunanen N, Lämsä A, Astola H, Räty T. Deepcloud - A Fully Convolutionnal Neural Network for Cloud and Shadow Masking in Optical Satellite Images. In 2018 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2018 - Proceedings. IEEE Institute of Electrical and Electronic Engineers . 2018. p. 2107-2110. 8517484 https://doi.org/10.1109/IGARSS.2018.8517484