Ulva prolifera detection with dual-polarization GF-3 SAR data

Abstract

Ulva prolifera is a marine hazard occurs frequently in the Yellow Sea of China, which has led to continuous impacts on the marine environment. Accurate and timely information on Ulva prolifera monitoring by satellite remote sensing, plays a significant role in the marine ecological environment and coastal tourism development. Compared with optical remote sensing, Synthetic Aperture Radar (SAR) has the capability of imaging the earth's surface independent of sunlight and cloud, as well as in high spatial resolution. Polarimetric Synthetic Aperture Radar (PolSAR) backscattered signal that contains scattering and phase information, has a great advantage in target observation. An experiment based on GF-3 dual polarimetric characteristics features is implemented, which indicates it is suitable for detecting Ulva prolifera. The detection results are extracted based on the proposed method in Lianyungang coastal zone. Compared with the ground truth of expert interpretation, the kappa coefficient is 0.94, and the overall accuracy is 0.99.