Effects of shadow correction on vegetation and land cover classification from high resolution aerial images

Teemu Kumpumäki, Tarmo Lipping

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

    1 Citation (Scopus)

    Abstract

    The effects of shadow correction on the classification of vegetation and land cover is studied in high resolution (10 × 10 cm) aerial images. Shadow detection reuses the feature set derived from the imagery for vegetation classification. A separate model is used to classify data in the first pass into three classes: water, land and shadows. Areas classified as shadow are then corrected using a regression based model and the shadow pixel features are recalculated and updated into the feature set. The results indicate that shadow correction can significantly improve classification results in vegetation mapping; in our case the classification accuracy increased from 35.1 to 46.1 in a shadowy test area.
    Original languageEnglish
    Title of host publication2016 IEEE International Geoscience and Remote Sensing Symposium (IGARSS)
    Subtitle of host publication10-15 July 2016, Beijing, China
    PublisherIEEE
    Pages751-754
    Number of pages4
    ISBN (Electronic)978-1-5090-3332-4
    DOIs
    Publication statusPublished - 2016
    Publication typeA4 Article in conference proceedings
    EventINTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM -
    Duration: 2 Dec 2016 → …

    Publication series

    Name
    ISSN (Electronic)2153-7003

    Conference

    ConferenceINTERNATIONAL GEOSCIENCE AND REMOTE SENSING SYMPOSIUM
    Period2/12/16 → …

    Publication forum classification

    • Publication forum level 1

    Fingerprint

    Dive into the research topics of 'Effects of shadow correction on vegetation and land cover classification from high resolution aerial images'. Together they form a unique fingerprint.

    Cite this