现代谱估计技术被用于雷达成像中来提高图像分辨率和目标特征提取，但其缺点是需耗时的二维谱峰搜索，且由于采用码赛克技术，算法性能受小样本数限制．文中将成像高分辨算法中的两维参数估计问题转化为两个独立的一维问题，然后利用ＥＳＰＲＩＴ方法或ＭＵＳＩＣ方法求得每维参数估计，这种方法无需两维谱峰搜索，算法性能在小样本时有很大改善，计算机仿真结果表明了文中算法的有效性． Modern spectral estimation techniques are used in radar imaging to improve image resolution and target feature extraction, but the disadvantage is that it takes time-consuming two-dimensional peak search, and the performance of the algorithm is limited by the small number of samples due to the use of the codec technique. In this paper, the two-dimensional parameter estimation problem in high-resolution imaging algorithm is transformed into two independent one-dimensional problems, and then the ESPRIT method or MUSIC method is used to obtain each dimension parameter estimation. This method does not need two-dimensional peak search. Small sample has greatly improved, the computer simulation results show the effectiveness of the algorithm in the paper.