利用海杂波的循环平稳特性,提出了一种基于奇异值分解的纹理分量抑制算法,并对其性能进行了分析。首先结合海杂波的循环平稳模型,推导了循环均值和循环自相关函数的封闭表达式。利用海杂波强度数据构造Hankle矩阵并进行奇异值分解,由于奇异值分解后得到的奇异值主要成分是与纹理分量谐波频率对应的特征量,而循环均值在纹理分量谐波频点处出现峰值,因此可以通过抑制奇异值主要成分实现循环均值的峰值抑制。奇异值抑制后,通过矩阵重构和数据恢复,最终实现纹理分量的抑制。最后,分析了Hankle矩阵维数对抑制性能的影响,并分别采用仿真数据和实测数据验证了算法的有效性。
A method for the suppression of texture component based on singular value decomposition is presented accord- ing to the cyclostationary property of sea clutter and its performance is analyzed. Firstly, the closed form expression for cy- clic mean and cyclic correlation function is derived with the cyclostationary model. Then, Hankle matrix is formed by using sea clutter intensity data, and the main components of the singular values, calculated by singular value decomposition, are the characteristic values corresponding with the harmonic frequencies of the texture component. In the meanwhile, the cyclic mean peaks at the harmonic frequencies of the texture component, so the peaks of the cyclic mean can be suppressed by suppressing the main components of singular values. After the suppression of singular values, texture component can be sup- pressed by matrix restructure and data restore. Finally, the influence of Hankle matrix dimension on the suppression capaci- ty is investigated and the effectiveness of the proposed method is verified with both simulated data and measured sea clutter data.