Mapping sub-pixel proportional land cover with AVHRR imagery

P M Atkinson, M. E. J. Cutler, H. Lewis

    Research output: Contribution to journalArticle

    229 Citations (Scopus)

    Abstract

    A problem with NOAA AVHRR imagery is that the intrinsic scale of spatial variation in land cover in the U.K. is usually finer than the scale of sampling imposed by the image pixels. The result is that most NOAA AVHRR pixels contain a mixture of land cover types (sub-pixel mixing). Three techniques for mapping the sub-pixel proportions of land cover classes in the New Forest, U.K. were compared: (i) artificial neural networks (ANN); (ii) mixture modelling; and (iii) fuzzy c-means classification. NOAA AVHRR imagery and SPOT HRV imagery, both for 28 June 1994, were obtained. The SPOT HRV images were classified using the maximum likelihood method, and used to derive the 'known' sub-pixel proportions of each land cover class for each NOAA AVHRR pixel. These data were then used to evaluate the predictions made (using the three techniques and the NOAA AVHRR imagery) in terms of the amount of information provided, the accuracy with which that information is provided, and the ease of implementation. The ANN was the most accurate technique, but its successful implementation depended on accurate co-registration and the availability of a training data set. Supervised fuzzy c-means classification was slightly more accurate than mixture modelling.

    Original languageEnglish
    Pages (from-to)917-935
    Number of pages19
    JournalInternational Journal of Remote Sensing
    Volume18
    Issue number4
    DOIs
    Publication statusPublished - 10 Mar 1997

    Cite this

    @article{b7f693e1f38046618f7eb6701d8bdf0f,
    title = "Mapping sub-pixel proportional land cover with AVHRR imagery",
    abstract = "A problem with NOAA AVHRR imagery is that the intrinsic scale of spatial variation in land cover in the U.K. is usually finer than the scale of sampling imposed by the image pixels. The result is that most NOAA AVHRR pixels contain a mixture of land cover types (sub-pixel mixing). Three techniques for mapping the sub-pixel proportions of land cover classes in the New Forest, U.K. were compared: (i) artificial neural networks (ANN); (ii) mixture modelling; and (iii) fuzzy c-means classification. NOAA AVHRR imagery and SPOT HRV imagery, both for 28 June 1994, were obtained. The SPOT HRV images were classified using the maximum likelihood method, and used to derive the 'known' sub-pixel proportions of each land cover class for each NOAA AVHRR pixel. These data were then used to evaluate the predictions made (using the three techniques and the NOAA AVHRR imagery) in terms of the amount of information provided, the accuracy with which that information is provided, and the ease of implementation. The ANN was the most accurate technique, but its successful implementation depended on accurate co-registration and the availability of a training data set. Supervised fuzzy c-means classification was slightly more accurate than mixture modelling.",
    author = "Atkinson, {P M} and Cutler, {M. E. J.} and H. Lewis",
    year = "1997",
    month = "3",
    day = "10",
    doi = "10.1080/014311697218836",
    language = "English",
    volume = "18",
    pages = "917--935",
    journal = "International Journal of Remote Sensing",
    issn = "0143-1161",
    publisher = "Taylor & Francis",
    number = "4",

    }

    Mapping sub-pixel proportional land cover with AVHRR imagery. / Atkinson, P M ; Cutler, M. E. J. ; Lewis, H.

    In: International Journal of Remote Sensing, Vol. 18, No. 4, 10.03.1997, p. 917-935.

    Research output: Contribution to journalArticle

    TY - JOUR

    T1 - Mapping sub-pixel proportional land cover with AVHRR imagery

    AU - Atkinson, P M

    AU - Cutler, M. E. J.

    AU - Lewis, H.

    PY - 1997/3/10

    Y1 - 1997/3/10

    N2 - A problem with NOAA AVHRR imagery is that the intrinsic scale of spatial variation in land cover in the U.K. is usually finer than the scale of sampling imposed by the image pixels. The result is that most NOAA AVHRR pixels contain a mixture of land cover types (sub-pixel mixing). Three techniques for mapping the sub-pixel proportions of land cover classes in the New Forest, U.K. were compared: (i) artificial neural networks (ANN); (ii) mixture modelling; and (iii) fuzzy c-means classification. NOAA AVHRR imagery and SPOT HRV imagery, both for 28 June 1994, were obtained. The SPOT HRV images were classified using the maximum likelihood method, and used to derive the 'known' sub-pixel proportions of each land cover class for each NOAA AVHRR pixel. These data were then used to evaluate the predictions made (using the three techniques and the NOAA AVHRR imagery) in terms of the amount of information provided, the accuracy with which that information is provided, and the ease of implementation. The ANN was the most accurate technique, but its successful implementation depended on accurate co-registration and the availability of a training data set. Supervised fuzzy c-means classification was slightly more accurate than mixture modelling.

    AB - A problem with NOAA AVHRR imagery is that the intrinsic scale of spatial variation in land cover in the U.K. is usually finer than the scale of sampling imposed by the image pixels. The result is that most NOAA AVHRR pixels contain a mixture of land cover types (sub-pixel mixing). Three techniques for mapping the sub-pixel proportions of land cover classes in the New Forest, U.K. were compared: (i) artificial neural networks (ANN); (ii) mixture modelling; and (iii) fuzzy c-means classification. NOAA AVHRR imagery and SPOT HRV imagery, both for 28 June 1994, were obtained. The SPOT HRV images were classified using the maximum likelihood method, and used to derive the 'known' sub-pixel proportions of each land cover class for each NOAA AVHRR pixel. These data were then used to evaluate the predictions made (using the three techniques and the NOAA AVHRR imagery) in terms of the amount of information provided, the accuracy with which that information is provided, and the ease of implementation. The ANN was the most accurate technique, but its successful implementation depended on accurate co-registration and the availability of a training data set. Supervised fuzzy c-means classification was slightly more accurate than mixture modelling.

    U2 - 10.1080/014311697218836

    DO - 10.1080/014311697218836

    M3 - Article

    VL - 18

    SP - 917

    EP - 935

    JO - International Journal of Remote Sensing

    JF - International Journal of Remote Sensing

    SN - 0143-1161

    IS - 4

    ER -