基于SAR子孔径图像的浅海水下地形探测

Shallow Sea Underwater Topography Detection Based on SAR Subaperture Image

摘要:

沿海地区水下地形的精确测量是人类开展海洋活动的关键，合成孔径雷达（SAR）为浅海地形的探测提供了一种新的手段，其中TerraSAR-X 的聚束模式以较长的积分时间得到高分辨率的 SAR 数据，从而能较为精确地反演浅海的海底地形。传统 SAR 图像水下地形探测基于波周期不变的假设，这不仅需要已知初始水深求解波周期，还给探测结果带来一定误差。本文提出一种基于子孔径图像的水下地形探测方法，将一景SAR图像分解成多景时间间隔固定的子孔径图像，利用子孔径图像间的时间间隔，求解不断变化的波周期，从而获得更加精确的水下地形。使用在海南蜈支洲岛的 TerraSAR-X 数据验证了此方法的可行性，将此方法反演得到的结果与GEBCO数据进行比较，发现两者吻合较好（MAE为2.8 m，MRE为23.91%），证明了此方法在浅海反演水深的巨大潜力。

Abstract:

Accurate measurement of underwater terrain in coastal areas is the key to human activities in the ocean. Synthetic aperture radar (SAR) provides a new means for the detection of shallow sea terrain. TerraSAR-X bunching mode can obtain high-resolution SAR data with a long integration time, so that the seabed terrain of shallow sea can be inverted more accurately. Traditional SAR image underwater terrain detection is based on the assumption of constant wave period, which not only needs to know the initial water depth to solve the wave period, but also brings some errors to the detection results. In this paper, an underwater terrain detection method based on subaperture images is proposed. A SAR image is decomposed into subaperture images with fixed multi-scene time intervals, and the time interval between subaperture images is used to solve the changing wave period, so as to obtain more accurate underwater terrain. The TerraSAR-X data of Wuzhizhou Island in Hainan Province was used to verify the feasibility of this method. The results obtained by inversion of this method were compared with the GEBCO data, and it was found that the two were in good agreement（the MAE was 2.8m and the MRE was 23.91%）. It is proved that this method has great potential for retrieving water depth in shallow sea.