To deal with the non-Caussian noise in standard 2-D SAR images, the deramped signal in imaging plane, and the possible symmetric distribution of complex noise, the fourth-order cumulant of complex process is introduce...To deal with the non-Caussian noise in standard 2-D SAR images, the deramped signal in imaging plane, and the possible symmetric distribution of complex noise, the fourth-order cumulant of complex process is introduced into SAR tomography. With the estimated AR parameters of ARMA model of noise through Yule-Walker equation, the signal series of height is pre-filtered. Then, through ESPRIT, the spectrum is obtained and the aperture in height direction is synthesized. Finally, the SAR tomography imaging of scene is achieved. The results of processing on signal with non-Gaussian noise demonstrate the robustness of the proposed method. The tomography imaging of the scenes shows that the higher-order spectrum analysis is feasible in the application.展开更多
非高斯性数据间的因果网络已经在经济学、生物学和环境学等学科得到了广泛应用.DirectLingam(Direct Method for Learning a Linear Non-Gaussian Structural Equation Model)算法是其中一个经典解法,但其存在维度达到25维度以上时外生...非高斯性数据间的因果网络已经在经济学、生物学和环境学等学科得到了广泛应用.DirectLingam(Direct Method for Learning a Linear Non-Gaussian Structural Equation Model)算法是其中一个经典解法,但其存在维度达到25维度以上时外生变量(exogenous variable)识别率低的问题,进而产生级联效应,使得整个网络的估计误差随着层数增大越来越大.为此提出了一种基于负熵局部选择外生变量的DirectLingam算法(LS-DirectLingam),把变量的非高斯性作为外生变量选择的标准,用负熵来度量变量的非高斯,选择负熵最大的k个变量存入局部目标变量集合Lv中,在集合Lv中进一步去寻找外生变量,从而提高了外生变量的识别率.与基本的DirectLingam算法进行实验比较,结果表明LS-DirectLingam算法优于DirectLingam算法.展开更多
基金supported partly by the New Century Excellent Talents in University(23901019)the Sichuan Provincial Youth Science and Technology Foundation(06ZQ026-006).
文摘To deal with the non-Caussian noise in standard 2-D SAR images, the deramped signal in imaging plane, and the possible symmetric distribution of complex noise, the fourth-order cumulant of complex process is introduced into SAR tomography. With the estimated AR parameters of ARMA model of noise through Yule-Walker equation, the signal series of height is pre-filtered. Then, through ESPRIT, the spectrum is obtained and the aperture in height direction is synthesized. Finally, the SAR tomography imaging of scene is achieved. The results of processing on signal with non-Gaussian noise demonstrate the robustness of the proposed method. The tomography imaging of the scenes shows that the higher-order spectrum analysis is feasible in the application.
文摘非高斯性数据间的因果网络已经在经济学、生物学和环境学等学科得到了广泛应用.DirectLingam(Direct Method for Learning a Linear Non-Gaussian Structural Equation Model)算法是其中一个经典解法,但其存在维度达到25维度以上时外生变量(exogenous variable)识别率低的问题,进而产生级联效应,使得整个网络的估计误差随着层数增大越来越大.为此提出了一种基于负熵局部选择外生变量的DirectLingam算法(LS-DirectLingam),把变量的非高斯性作为外生变量选择的标准,用负熵来度量变量的非高斯,选择负熵最大的k个变量存入局部目标变量集合Lv中,在集合Lv中进一步去寻找外生变量,从而提高了外生变量的识别率.与基本的DirectLingam算法进行实验比较,结果表明LS-DirectLingam算法优于DirectLingam算法.
