Rice-Planted Area Mapping Using Small Sets of Multi-Temporal SAR Data

A rice-planted area map is a basic information resource for rice production management. Synthetic aperture radar (SAR) is an appropriate technique for rice mapping and so far is mostly based on extracting time series changes of backscattering (sigma(0)) in a rice-planted area. However, sometimes there is not enough data to extract the sigma(0) curve for the area. To overcome this problem of a lack of data, we propose a method to detect rice-planted area by using small sets of multi-temporal SAR data. This method also addresses the fluctuation of sigma(0) values between SAR measurements. We have applied the method using multi-temporal ALOS/PALSAR data acquired over five years during the dry season. The rice-planted area was well detected and the viability of this method was demonstrated.