RBF插值在SAR距离弯曲校正中的应用

Application of RBF Interpolation in Range Migration Correction of SAR

在SAR成像算法中，距离徙动会使距离向和方位向发生耦合，成像质量下降。距离-多普勒算法中，通过插值补偿距离徙动量。文章在分析三种典型插值算法的基础上，提出了利用径向基网络作为函数逼近的插值校正方法。因为径向基网络具有最佳逼近任一函数的性质，且逼近的函数具有很好的光滑性和收敛性，所以得出了径向基网络插值校正效果优于典型插值的结论，并对这几种插值算法进行了仿真，通过比较成像结果和成像质量指标，进一步验证了结论的正确性。

In SAR Imaging algorithm, range migration results in the coupling between range and azimuth, and affects the imaging quality. Range-Doppler method carries out Range migration correction （RMC） by interpolation in range-doppler domain. In this paper, three typical interpolations are proposed, and a new interpolation algorithm which uses the RBF function approximation is presented for RMC. The RBF network can greatly approximate any function which is smooth and convergent. Therefore, it shows that a better imaging quality can be derived than other typical interpolation methods, and comparing the imaging results after correction with the corresponding quality index, we verify the effectiveness of this method.