论文部分内容阅读
由于基于l1范数的压缩感知理论模型无法充分挖掘信号的稀疏性,因此在重建过程中无法实现对待重构系数的等权值约束,进而导致在信噪比较低时,噪声分布的不稀疏性会严重影响目标信息的重建,造成成像结果中会出现大量虚假目标,成像性能急剧下降。本文在深入分析了加权l1范数模型的基础上,提出了一种更加稳健的适用于含噪模式下的高分辨率压缩感知微波成像模型。该模型在借鉴常规加权l1范数模型的基础上,针对权重选择及加权方式进行了修正,使得权值的变化程度和权值大小分离,可以做到相同的惩罚约束,从而实现成像过程中噪声分量的有效抑制,实验结果说明了低信噪比下所提模型的有效性。
As the compressible sensing theoretical model based on the l1 norm can not fully exploit the signal sparsity, the equal weight constraint on the reconstructed coefficients can not be achieved in the reconstruction process, which leads to the non-sparse distribution of the noise when the signal-to-noise ratio is low Sex will seriously affect the reconstruction of the target information, resulting in a large number of false imaging results will result in a sharp decline in imaging performance. Based on the in-depth analysis of the weighted l1 norm model, this paper proposes a more robust high-resolution compressive-sensing microwave imaging model suitable for the noisy mode. Based on the conventional weighted l1 norm model, the model is modified for weight selection and weighting to separate the degree of weight change from the weight value, so that the same penalty constraint can be achieved and the noise The effective suppression of the component, the experimental results show the effectiveness of the proposed model under low signal-to-noise ratio.