Single and Multi-temporal Filtering Comparison on Synthetic Aperture Radar Data for Agriculture Area Classification

The agricultural area has the most dynamic characteristic between the beginning of planting season and harvesting season, which often monitored by remote sensing data. Synthetic Aperture Radar (SAR) is an active remote sensing provides cloud free data that is suitable for tropical country such as Indonesia. One of the main and critical preprocessing of SAR data is speckle noise filtering. However, few studies have evaluated the effect on single and multi-temporal filtering in the agricultural area, especially ALOS2/PALSAR2 data. This study was performed to compare the single and multi-temporal speckle noise filtering characteristics in temporal data of ALOS2/PALSAR2 and the effects of both filtering in agriculture area classification. Overall, the single filtering image performed higher accuracy compared to the multi-temporal filtering image. The multi-temporal filtering image has sharper and clearer edge compared to the single filtering image. However, it also contains more salt and pepper noise than the single filtering image. This study showed that the single filtering is better to be used in temporal data of ALOS2/PALSAR2 for the agricultural area that has temporal change.