Multitemporal Sar Image Despeckling Based on Immse Filtering
Mohamed Yahia, Tarig Ali, Md Maruf Mortula, Samy El Mahdi, Nuwanthi Sashipraba Arampola · 2020
Most multitemporal filtering techniques assume that pixels remain unchanged over the time so that all the pixels at the same position in the time series are averaged. However, this assumption is not suitable in a long-period time series. To overcome this problem, a novel approach for temporal filtering of SAR image time series by applying the iterative minimum mean square error (IMMSE) approach is proposed. Real and simulated time series data sets were tested. Results showed that the IMMSE preserved the temporal unchanged areas filtered and enhanced the blurred temporal changed details.