Rice Recognition Using Multi-temporal and Dual Polarized Synthetic Aperture Radar Images

Synthetic aperture radar has been used widely in crop recognition and monitoring. In the past, most of the studies of rice mapping and yield estimation are based on single polarized SAR data. In order to evaluate the feasibility of rice recognition approach and compress the amount of SAR data, multi-temporal and dual polarized SAR data were used here. In rice growing stage, two dates and dual polarization (VV & HH) ENVISAT ASAR data were acquired in Changle study area. With the analysis of rice and other land cover classes backscatter coefficient, an image processing method of difference index was applied to enhance rice information and decreased the confusion between rice and non-rice. Threshold technique was carried out finally for extracting rice information. Results indicate that the rice information extraction achieves a 91.02% overall accuracy. Therefore, the method is an effective one for SAR information recognition in distinguishing rice and other land cover categories.