Automatic Approach to VHR Satellite Image Classification
PBN-AR
Instytucja
Wydział Geodezji i Kartografii (Politechnika Warszawska)
Źródłowe zdarzenia ewaluacyjne
Informacje podstawowe
Główny język publikacji
en
Czasopismo
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
ISSN
1682-1750
EISSN
Wydawca
International Society for Photogrammetry and Remote Sensing
DOI
Rok publikacji
2016
Numer zeszytu
XLI-B7
Strony od-do
277-282
Numer tomu
Liczba arkuszy
0.5
Słowa kluczowe
en
remote sensing, classification, texture analysis, granulometry, VHR, satellite image, vegetation indices
Konferencja
Indeksowana w Scopus
nie
Indeksowana w Web of Science Core Collection
tak
Liczba cytowań z Web of Science Core Collection
Nazwa konferencji (skrócona)
XXIII ISPRS Congress
Nazwa konferencji
XXIII International Society for Photogrammetry and Remote Sensing Congress
Początek konferencji
2016-07-12
Koniec konferencji
2016-07-19
Lokalizacja konferencji
Prague
Kraj konferencji
CZ
Lista innych baz czasopism i abstraktów w których była indeksowana
Streszczenia
Język
en
Treść
In this paper, we present a proposition of a fully automatic classification of VHR satellite images. Unlike the most widespread approaches: supervised classification, which requires prior defining of class signatures, or unsupervised classification, which must be followed by an interpretation of its results, the proposed method requires no human intervention except for the setting of the initial parameters. The presented approach bases on both spectral and textural analysis of the image and consists of 3 steps. The first step, the analysis of spectral data, relies on NDVI values. Its purpose is to distinguish between basic classes, such as water, vegetation and non-vegetation, which all differ significantly spectrally, thus they can be easily extracted basing on spectral analysis. The second step relies on granulometric maps. These are the product of local granulometric analysis of an image and present information on the texture of each pixel neighbourhood, depending on the texture grain. The purpose of texture analysis is to distinguish between different classes, spectrally similar, but yet of different texture, e.g. bare soil from a built-up area, or low vegetation from a wooded area. Due to the use of granulometric analysis, based on mathematical morphology opening and closing, the results are resistant to the border effect (qualifying borders of objects in an image as spaces of high texture), which affect other methods of texture analysis like GLCM statistics or fractal analysis. Therefore, the effectiveness of the analysis is relatively high. Several indices based on values of different granulometric maps have been developed to simplify the extraction of classes of different texture. The third and final step of the process relies on a vegetation index, based on near infrared and blue bands. Its purpose is to correct partially misclassified pixels. All the indices used in the classification model developed relate to reflectance values, so the preliminary step of recalculation of pixel DNs to reflectance is required. Thanks to this, the proposed approach is in theory universal, and might be applied to different satellite system images of different acquisition dates. The test data consists of 3 Pleiades images captured on different dates. Research allowed to determine optimal indices values. Using the same parameters, we obtained a very good accuracy of extraction of 5 land cover/use classes: water, low vegetation, bare soil, wooded area and built-up area in all the test images (kappa from 87% to 96%). What constitutes important, even significant changes in parameter values, did not cause a significant declination of classification accuracy, which demonstrates how robust the proposed method is.
Inne
System-identifier
WUTb1bddc31e5084509b1ea1914f0c6a956
CrossrefMetadata from Crossref logo
Cytowania
Liczba prac cytujących tę pracę
Brak danych
Referencje
Liczba prac cytowanych przez tę pracę
Brak danych