利用FB-MAPES算法估计Tomography SAR高度维信号源数

层析成像合成孔径雷达(Tomography SAR)结合空间谱估计技术解决层叠效应时，首先要估计出高度维信号源数目，信号源数估计得正确与否直接关系到后续处理的估计性能。用传统的信息论方法估计时，由于数据受乘性噪声影响，估计性能下降。该文提出一种新的信号源数估计方法:双向平滑多视幅相估计(FB-MAPES)滤波器算法。通过Monte Carlo仿真对信息论方法中的GMDL算法和FB-MAPES方法的估计性能进行了比较，试验结果表明该文提出的方法能够更好地适应乘性噪声的影响，因而在估计Tomography SAR高度维信号源数时性能优于传统的信息论方法。

On Detection of Number of Tomography SAR Signals in the Elevation Direction Using the FB-MAPES Method

When using Tomography SAR combined with spatial spectrum estimation to solve layover effect, the number of signals should be estimated in the elevation direction first. The accuracy of number of signals is directly related to the following estimation performance. The conventional Information Theoretic Criteria (ITC) methods are usually influenced by multiplicative noise of the data, consequently, the performance of the estimation decreases. In this paper, a new method, FB-MAPES, is proposed to detect the number of signals. Then the performance of General Minimum Descriplion Length (GMDL) and FB-MAPES is compared via the Monte Carlo simulation. The experimental results confirm the validity of the FB-MAPES method. Moreover, the proposed method is prior to conventional ITC in detecting the number of tomography SAR signals in the elevation direction.