Mapping of Spatio-Temporal Changes of Surface Water Using Sentinel-1 SAR Images

Exploration of surface water bodies and their spatial definition has great importance in water resources management and helps understand hydrological processes in the country. Fast, regular, and effective ways of mapping watercourses and their surroundings through remote sensing methods are crucial tools for capturing change and predicting hazards. The last decades have brought new data products, open-source software, and evaluation procedures that allow low-cost mapping of surface water objects. A widespread and sought-after option for this type of mapping is the use of SAR (Synthetic Aperture Radar) image products. Just through SAR technology that it is possible to identify changes in water in a relatively short time and at the same time under any meteorological conditions thanks to backscattered microwave radiation. This paper presents the possibilities of using SAR technology and its data for long-term temporal mapping of meteorological-hydrological changes in surface water using satellite images of Sentinel-1 product level GRD. As regards surface water extraction, this process is performing by segmenting the threshold values according to the Otsu principle. The water surfaces are then interpreted into the form of water masks of objects by the binarization of the final image. These values are subsequently compared with the supervised classification RFC (Random Forest Classifier) method results. Suitable processing and evaluation procedures conclude that the more suitable polarization configuration for mapping water bodies is VV (vertical-vertical) polarization. As for speckle filter tools to eliminate radar noise, the most suitable option seems to be using a Lee filter. To achieve more accurate results of the extracted water bodies, it is then appropriate to implement quantitative statistical indicators of accuracy and their numerical interpretation of the reliability of results. This paper aims to demonstrate the advantages of using satellite SAR images for spatiotemporal mapping of surface water in the landscape to observe hydrological processes due to inundation, ecological and meteorological changes, and anthropogenic activity.

Evaluation of the change in synthetic aperture radar imaging using transfer learning and residual network

Change detection from synthetic aperture radar images becomes a key technique to detect change area related to some phenomenon as flood and deformation of the earth surface. This paper proposes a transfer learning and Residual Network with 18 layers (ResNet-18) architecture-based method for change detection from two synthetic aperture radar images. Before the application of the proposed technique, batch denoising using convolutional neural network is applied to the two input synthetic aperture radar image for speckle noise reduction. To validate the performance of the proposed method, three known synthetic aperture radar datasets (Ottawa; Mexican and for Taiwan Shimen datasets) are exploited in this paper. The use of these datasets is important because the ground truth is known, and this can be considered as the use of numerical simulation. The detected change image obtained by the proposed method is compared using two image metrics. The first metric is image quality index that measures the similarity ratio between the obtained image and the image of the ground truth, the second metrics is edge preservation index, it measures the performance of the method to preserve edges. Finally, the method is applied to determine the changed area using two Sentinel 1 B synthetic aperture radar images of Eddahbi dam situated in Morocco.