基于墨西哥帽小波变换的机载SAR海面风场反演

Ocean surface wind field retrieval from airborne SAR imagery based on Mexican-hat wavelet transform

针对从机载SAR探测图像中反演风向中如何提取图像中的风条纹信息的问题,给出基于墨西哥帽小波的风条纹信息提取方法,结合二维快速Fourier变换反演海面风向。以中国南海海域一组机载SAR实测数据为例,采用该方法反演风向,并运用CMOD4、CMOD-IFR2和CMOD5模式反演风速,验证海面风场反演方法特性。结果表明,墨西哥帽小波法较好地解决了机载SAR探测图像的风条纹信息提取问题,反演精度优于局部梯度法,3种风速反演模式均可用于机载SAR探测数据风速反演。

How to extract wind streaks information from imagery is the hard point in the retrieval of wind direction from airborne SAR sounding imagery. The wind streaks information extraction method based on Mexican-hat wavelet combining with two-dimensional fast fourier transform was proposed to solve this problem. Using C-band empirical models such as CMOD4, CMOD-IFR2 and CMOD5, associated with the proposed wind direction retrieval method, the wind field was obtained from airborne SAR sounding data. The characteristics of ocean surface wind field retrieval methods were verified by retrieving wind filed from a group of airborne SAR sounding data of South China Sea. The results show that the Mexican-hat wavelet algorithm solves the problem of extracting wind streaks information from airborne SAR sounding imagery, and the retrieval accuracy is better than that of the local-gradient algorithm, while three kinds of wind speed retrieval models can be used to retrieve wind speed from airborne SAR sounding data.