Tropical Cyclone Rainbands in SAR Images

A tropical cyclone is a natural disaster that occurs frequently and usually brings much precipitation when it makes landfall. Tropical cyclone rainbands are an essential element in the tropical cyclone system and are often associated with tropical cyclone’s rainfall. Synthetic aperture radar (SAR) and satellite cloud images are two main observation methods to obtain tropical cyclone images. However, most previous studies focused on the satellite cloud images of tropical cyclones, and few studies on the tropical cyclones in SAR images, especially the rainbands. Therefore, in this paper, we collected SAR images containing tropical cyclone rainbands elements to observe the performance of rainbands in SAR images. Based on these data, the rainbands region in SAR images is identified and labeled by human eyes. This paper establishes a neural network data set containing tropical cyclone rainbands features to automatically extract tropical cyclone rainbands region from SAR images by using a neural network. Finally, one independent SAR image of tropical cyclones was selected for testing, and the extracted results showed great consistency with the human visual interpretation results.