Wykorzystanie metod geostatystyki do wspomagania klasyfikacji mikrofalowych zdjęć satelitarnych

• Autorzy: Stankiewicz. K. , Wiśniewska E.

Warianty tytułu

EN

The use of geostatistics to assist classification of microwave satellite images

• Języki publikacji: PL

Abstrakty

EN

Microwave images from ENVISAT satellite can be used for the recognition of land cover classes. They can be extremely important on the occasions when such recognition should be done at a specified moment but optical satellite images are unavailable due to persisting overcast. However, the application of microwave images to land cover classification requires special handling, because the registered microwave signal depends on various factors which make the unique interpretation of images a complex task. The presented paper is focused on vegetation classes representing short natural vegetation and agricultural crops. The microwave backscattering from vegetation depends strongly on canopy architecture as well as on water content in plants and in soil. The spatial variability of moisture disturbs the visual interpretation of microwave images and makes their automatic classification difficult. In order to elaborate efficient and robust classification methods, the spatial variability of signatures referring to land cover classes has to be analyzed in the first step. The analysis of signature variability for vegetation classes was presented using ENVISAT ASAR microwave images acquired during vegetation growth season in 2005. Our test site is located in Wielkopolska in the vicinity of Dezydery Chłapowski Agro-ecological Landscape Park. It is a rural area with prevailing agriculture land use. Beside arable land, orchards and plantations there are the following other significant land cover classes: deciduous and coniferous forest, grasslands, urban area and water bodies. The signatures of vegetation classes were investigated considering the date and the parameters of images registration as well as various characteristics of the test site area such as spatial variability of biomass, moisture content in plants and in soil. Several information layers were considered in the project in order to characterize the investigated area: digital elevation model, soil maps, topographic maps, satellite images acquired in the optical range, the results of point measurements of soil moisture and biomass. Point measurements of volumetric soil moisture (VSM) taken in the upper layer of soil were interpolated in order to estimate the spatial distribution of VSM over the whole area of the agricultural field. Arc Map software and Geostatistical Analyst module were used for the geostatistical analysis of the experimental data. The empirical semivariogram calculated from measured VSM data was investigated. The ordinary kriging was applied in order to estimate soil moisture over the field. The interpolation results were compared with the spatial distribution of NDVI calculated from satellite images acquired in the optical range on various dates. This comparison shows that spatial distribution of soil moisture is in agreement with some stable environmental features. This observation can help to identify areas, which can be critical to the accuracy of vegetation recognition on microwave images. The correlation of microwave backscattering with the estimated spatial distribution of soil moisture was also investigated. The analysis shows that cross-polarized images acquired with VH or HV polarization are better suited to vegetation classification than co-polarized ones because they are less sensitive to moisture variability.

Słowa kluczowe

PL

klasyfikacja; ENVISAT; polaryzacja; wilgotność gleby; kriging

EN

classification; ENVISAT; polarization; soil moisture; kriging

• Wydawca: Polskie Towarzystwo Informacji Przestrzennej
• Czasopismo: Roczniki Geomatyki
• Rocznik: 2006
• Tom: T. 4, z. 3
• Strony: 191--201
• Opis fizyczny: Bibliogr. 11 poz.

Twórcy

autor

Stankiewicz. K.

autor

Wiśniewska E.

• Instytut Geodezji i Kartografii,

Bibliografia

• 1. Notarnicola C., Angiulli M., Posa F., 2006: Use of Radar and Optical Remotely Sensed Data for Soil Moisture Retrieval Over Vegetated Areas, IEEE Trans. Geosci. Remote Sens., vol.44, no.4 (2006), 925-935.
• 2. eCognition User Guide 2002: Definiens Imaging GmbH, München
• 3. Goovaerts P., 1997: Geostatistics for Natural Resources Evaluation, Oxford University Press, Oxford – New York.
• 4. Johnston K., Ver Hoef J.M., Krivoruchko K., Lucas N., 1997: Using ArcGIS Geostatistical Analyst, ESRI, USA.
• 5. Mizgajski A., 1986: Niektóre uwarunkowania przepuszczalności warstwy przypowierzchniowej w rolniczo użytkowanych geokompleksach młodoglacjalnych, Badania Fizjograficzne nad Polską Zachodnią, t. XXXVI, seria A, Geografia fizyczna, PWN, Poznań-Warszawa, str. 137-154.
• 6. Oliver C., Quegan S., 1998: Understanding Synthetic Aperture Radar Images, Artech House, London, 1998.
• 7. Stankiewicz K.A., 2006: The Efficiency of Crop recognition on ENVISAT ASAR Images in Two Growing Seasons, IEEE Trans. Geosci. Remote Sens., vol. 44, no. 4 (2006), 806-814.
• 8. Tso B., Mather P.M., 2001: Classification Methods for Remotely Sensed Data, Taylor & Francis, London.
• 9. Usowicz B., Hajnos, M., Sokołowska Z., Józefaciuk G., Bowanko G., Kossowski J., 2004: Przestrzenna zmienność fizycznych i chemicznych właściwości gleby w skali pola i gminy, Acta Agrophysica, Rozprawy i monografie, Lublin.
• 10. Usowicz B., Usowicz Ł., 2004: Punktowe pomiary wilgotności gleby a jej przestrzenny rozkład na polach uprawnych, Acta Agrophysica, 4(2), 573-588.
• 11. Vecchia A.D., Ferrazzoli P., Guerriero L., Defourny P., Dente L., Mattia F., Satalino G., Strozzi T., Wegmüller U., 2006: Influence of Geometrical Factors on Crop Backscattering at C-Band, IEEE Trans. Geosci. Remote Sens., vol. 44, no. 4, 778-790.