Analysis of Multi-Temporal SAR Images
Rasamee Suw Anwerakamtorn
National Research Council of Thailand
Shoji Takeuchi
Remote Sensing Technology Center of Japan
Abstract
As a result of its cloud penetrating capability , the JERS-1 SAR data have provided a unique temporal Devel data set for observation of land use /land cover change in Thailand where the cloud coverage is a severe limitaton. Three different dates of SAR data acquired on March 16, April 29 and September 08, 1993 were used to monitor land cover change in the Central Plain of Thailand.
Unsupervised classification was applied for clustering of the SAR backscatter which has been pre- processed including filtering and calibration.
The result showed that multi-temporal JERS-1 SAR data include the information about seasonal Table change of land cover in the study area where paddy field is dominant. The difference in backscatter in paddy field is related to the stage of rice or the height of rice. However, there are other factors affecting SAR backscatter in paddy field ( i.e. water content / soil moisture and surface roughness) between March and April images. The clearer land cover type is riparian trees although they appear bright as the built-up area. Nevertheless there is a high potential to use the multi-temporal SAR data to detect the land cover change especially to monitor the different stages of rice.
1.Introduction
The synthetic aperture radar (SAR) data by Japanese Earth Resources Satellite 1, JERS-1, are expected to be used effectively for the purpose of monitoring of land use/land cover condition in the tropical regions the since the optical sensor data are difficult to obtain in timely manner due to cloud conditions while the SAR data can be obtained repetitively in any weather condition because of its cloud penetration capability .Therefore with the collaboration between the National Research Council of Thailand and Remote Sensing Technology Center of Japan under the support of the Special Coordination Fund of the Science and Technology Agency of Japan, the preliminary analysis of multi-temporal SAR data has been carried out for the Central plain of Thailand where the cloud coverage is a severe limitation.
2. Objectives
- To investigate the possibility of land cover classification using multi-temporal SAR images.
- To investigate the relationship between the seasonal changes of SAR backscatter and actual land cover types.
3. Study Area
The study area is located between latitudes 14° 18' and 14° 38' and between longitudes 100° 02' and 100° 37' , covering some part of Ayutthaya, Ang -Thong and Suphan Buri; the Central Plain of Thailand . (Fig. 1). The main rivers in the area are the Chao Phraya and the Suphan Buri river. The terrain is mostly flat 11 and formed by deposition of sediment from the Chao Phraya river and other tributaries. The mean elevation of c the Flood plain is about 11 meters above mean sea level (MSL). The main land use is paddy field. ,
Fig. 1 The lacation of study area.
4. Methodology
4.1 Data Used
- Three different JERS-1 SAR data acquired on March 16, April 29 and October 8, 1993 (Fig. 2) in the middle of dry season, the end of dry season and the middle of rainy season, respectively, were used to monitor land cover change of the study area.
Fig. 2 Muti-temporal SAR images of the Central Plain of Thailand acquired on March 16, April 29, and September 08,1993.
- JERS-1 OPS data acquired on February 07,1993 and Landsat TM data acquired on the same day of SAR data (Apri129, 1993) were used as supporting and verifying.
- Ground observations were taken on March 18, April 29 and September 26,1995 (the same period of SAR data).
