一种基于SVM的超分辨成像算法研究

A super-resolution algorithm for radar imaging based on SVM

为了提高逆合成孔径雷达(ISAR)成像分辨力,提出了一种基于支持向量机(SVM)的ISAR成像算法,通过适当地选择核函数的参数,可以在小观测角度的情况下,获得大观测角度图像,提高了ISAR成像的分辨力,实现了ISAR的超分辨成像。引入图像质量指数,仿真结果表明:基于SVM的超分辨方法在小转角条件下得到的目标像,具有和大转角实测数据下的目标像一样的图像质量,提出的方法可靠有效。

A new method is propsed to improve the resolution of ISAR image in this paper . The method uses the support vector machines based on radial basis kernel function（RBF）. By choosing the model parameter properly, a super-resolution image can be obtained under the condition of small azimuth angle. Using the image quality index, the feasibility and validity of super-resolution algorithm are testified by computer simulation.