基于噪声特性分析的稀疏度估计方法 Sparsity Estimation Method Based on Noise Characteristic Analysis

稀疏表示方法经过几十年的发展，在很多领域得到了深入研究与应用。信号稀疏度的估计是对信号进行稀疏分解的工作中的重要环节。根据噪声的特性，提出了一种稀疏度估计方法。即通过构造傅立叶基字典，利用噪声能量在全频域均匀分布的特点，通过遍历信号在全频谱上的稀疏特性，逐步确定准确的信号稀疏度。仿真实验结果表明，提出的方法能有效的完成对信号稀疏度的估计。 The sparse representation, after decades of development, has been deeply studied and applied in many fields. Estimation of signal sparsity is an important part of the work in the sparse decomposition of signal. According to the noise characteristics, this paper proposes a method for estimating sparse degree. By constructing the Fourier base dictionary, using the uniform distribution characteristics of noise energy in the whole frequency domain, and through the sparse characteristics of signal traversing the full spectrum, the accuracy of the signal sparsity is gradually determined. Simulation results show that the proposed method can effectively complete the estimation of signal sparsity.