The study of land use/land cover in coastal area of Chanthaburi province,Thailand, Using Cosmo-SkyMed : X band and Radarsat-2 :C-band

摘要

This paper presents the application of Synthetic Aperture Radar (SAR) for land use and land cover in coastal area.The objective is to study satellite data COSMO - SkyMed: X band and Radarsat-2: C band land use / land cover in coastal areas of Chanthaburi province.The important advantage of SAR imagery is that it is independent of the impact of climate factors, e.g., rain, haze, and cloud. As a result, the SAR imagery can be acquired from all seasons; therefore, it is suitable for tracking land use and land cover changes. Additionally, the potential capability of SAR sensor is to provide the backscattering data and these data can be performed to classify land use/land cover. However, the backscattering characteristics depend on the wavelength, direction of the radiated electromagnetic field, surface roughness, and moisture content. Cosmo-SkyMed (X-band, HH polarization) and Radarsat-2 (C band, HH,HV polarization) were used in this study. Those multi-temporal satellite images were acquired in May, August, and November 2013. The study area is in Chanthaburi province, Thailand. Land use in this area are paddy fields, cassava, rubber Trees, shrimp farms, mangrove, forest, orchard, buildings, forest, and water. The result of this study exhibited that the backscattering from buildings showed very high values in all wavelength and direction, where as the back scattering of water and shrimp farms were very low, which appears black on images because such surfaces are smooth forest, orchard and rubber Trees represented similar backscattering levels due to the characteristics of sparse vegetation canopy and paddy fields are medium roughness. The backscattering gray scale in this study area demonstrated different values from different materials according to the wavelength and direction of the radiated electromagnetic. This study found that surface roughness and moisture content are also the major factors of different backscattering values. If surfaces are rough, the brightness value will be white, such as roof and concrete road. Other important factors are moisture, wind, and incident angle. In addition, the result indicated that the major land use/land cover of this area, included paddy fields, shrimp farms, orchard, rubber trees,and cassava,respectively. Land uses/land cover within 15 km. of coastline in Chanthaburi province were also illustrated in this study. The accuracy of classification using the Radarsat-2 satellite data are 80% and using the Cosmo-SkyMedsatellite data is 75%.